See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/236102081

Development of a new chewing function questionnaire for assessment of a

self-perceived chewing function

Article in Community Dentistry And Oral Epidemiology · April 2013

DOI: 10.1111/cdoe.12048 · Source: PubMed

CITATIONS READS

37 952

4 authors:

Sanja Persic Antonija Palac

University of Zagreb 6 PUBLICATIONS 56 CITATIONS
45 PUBLICATIONS 548 CITATIONS
SEE PROFILE
SEE PROFILE

Tomislav Bunjevac Asja Celebić

University of Zagreb University of Zagreb
12 PUBLICATIONS 406 CITATIONS 177 PUBLICATIONS 3,439 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Asja Celebić on 03 January 2018.

The user has requested enhancement of the downloaded file.

Community Dent Oral Epidemiol 2013; 41; 565–573 Ó 2013 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
All rights reserved

Development of a new chewing Sanja Persic1, Antonija Palac2,


Tomislav Bunjevac3 and Asja Celebi
1
c4
Department of Prosthodontics, School of

function questionnaire for Dental Medicine, University of Zagreb,

Zagreb, Croatia, 2Department of
Prosthodontics, School of Medicine, Study of

assessment of a self-perceived
Dental Medicine, University of Split, Split,
Croatia, 3Department of Psychology, Faculty
of Humanities and Social Sciences,
University of Zagreb, Zagreb, Croatia,

chewing function 4
Department of Prosthodontics, School of
Dental Medicine and Clinical hospital
centre, University of Zagreb, Zagreb, Croatia


Persic S, Palac A, Bunjevac T, Celebi
c A. Development of a new chewing function
questionnaire for assessment of a self-perceived chewing function. Community
Dent Oral Epidemiol 2013; 41: 565–573. © 2013 John Wiley & Sons A/S. Published
by John Wiley & Sons Ltd

Abstract – Objective: To develop a new unidimensional questionnaire for

patients’ self-assessment of their chewing function (chewing function
questionnaire: CFQ) and to test its psychometric properties. Methods: A focus
group of 7 dentists and 15 prosthodontic patients generated a pool of 30
relevant items. A 5-point Likert scale (0–4) was used. A focus group reduced
some redundant items, and 21 items remained for pilot testing after which
additional four items were eliminated. Then, the 17-item questionnaire was
tested on a larger sample (200 participants). Factor analysis was obtained to
investigate the dimensionality of the questionnaire. Finally the 10 items
remained, and the CFQ was unidimensional. To test the psychometric
properties, the CFQ was administrated to 224 individuals: 100 patients were
already removable denture wearers (RDWs), 24 patients needed a removable
prosthodontic treatment, and 100 individuals had natural teeth (NT group).
Discriminative validity was tested between the RDWs and the NT groups.
Convergent validity and internal consistency were tested on 200 subjects. The
internal consistency was assessed by calculating the Cronbach’s alpha
coefficient and the average interitem correlation. The test–retest reliability was
tested on 60 subjects. Responsiveness was tested on 24 patients who received
new removable prosthodontic treatment. Results: The initial factor analysis
revealed that the items had been grouped in three different dimensions. By
eliminating items with complexities greater than one, additional six questions
were eliminated. One more item with the lowest correlation was further
eliminated. The final questionnaire included 10 items. Discriminative validity
showed significant differences between the NT group and the RDWs, as
predicted (P < 0.001). Convergent validity was confirmed by the Spearman’s
rank correlation. A significant positive association (P < 0.001) was found
between the CFQ and one general question about chewing difficulties, as well
as between the CFQ and five items from the OHIP49 related to chewing. The
test–retest reliability showed high intraclass correlation and no significant Key words: chewing; psychometric
differences between the two administrations of the CFQ (P > 0.05). The internal properties; questionnaire
consistency showed satisfactory Cronbach’s alpha values (0.916 for all subjects, Sanja Persic, Department of Prosthodontics,
0.742 for the NT group, and 0.852 for the RDWs). Responsiveness was School of Dental Medicine, University of
Zagreb, Gunduliceva 5, 10000 Zagreb,
confirmed by a significant difference between the baseline and the follow-up
Croatia
score (P < 0.001) and a high effect size (0.95). Conclusions: The developed CFQ Tel.: +385 1 4802211
provides an assessment instrument of the hypothetical construct invoked to Fax: +385 1 4802159
explain the chewing function. Psychometric properties are satisfactory e-mail: persic@sfzg.hr
demonstrating that the instrument is suitable for the assessment of a self- Submitted 28 May 2012;
perceived chewing function. accepted 8 March 2013

doi: 10.1111/cdoe.12048
565
Persic et al.
Patients’ attitude toward dental therapy has
become very important over the last 10 years and
should be acknowledged in dental treatment deci-
sions (1). A number of different questionnaires
have been developed in dentistry to express how
an individual perceives his/her dental appearance
(Orofacial Esthetic Scale), oral pain or discomfort
(Oral Health Impact Profile, Geriatric Oral Health
Assessment Index), jaw function limitations (Jaw
Functional Limitation Scale), questionnaire how an
individual perceives difficulties during mandibu-
lar functions (Mandibular Function Impairment
Questionnaire), etc. (2–6). Also, questionnaires for
the assessment of how oral problems affect the
quality of life have been developed (5, 7). Some of
these questionnaires are multidimensional, such as
OHIP (5), the Jaw Functional Limitation Scale
(JFLS) (2), General Oral Health Assessment Index
(GOHAI) (4), and Mandibular Function Impair-
ment Questionnaire (MFIQ) (6), while others are
unidimensional, such as Orofacial Esthetic Scale
(OES), which has been developed recently for the
measurement of orofacial esthetics (3).
Chewing function is one of the most important
functions of orofacial system. It has been reported
that an impairment of chewing ability showed
consistent association with an individual’s oral
health-related quality of life, number of teeth in
oral cavity, as well as an association with an indi-
vidual’s general health (8). Chewing function can
be assessed either by chewing tests (chewing dif-
ferent kinds of food or material and measuring
comminuted particles after sieving) and/or by a
questionnaire. Different chewing tests have been
described in the literature, for example, classic
sieving tests (9, 10) measuring the degree of differ-
ent food-comminuted particles or tests for the mea-
surement of particle size of a chewed wax cube or
of a high-viscosity silicone cubes (11, 12). Such tests
require specific and expensive equipment and are
not convenient for wider use, while questionnaires
are much easier and simpler methods for the
assessment of the chewing function. In a number
of studies, different questions have been used for
the assessment of the chewing function (13–16), for
example, individuals had to assess on a visual or
visual analogue scale their ability to chew different
kinds of food. Limitations of these questionnaires
are the lack of validation and the lack of psycho-
metric properties. Psychometric properties of a
questionnaire show how well the questionnaire
measures what is meant to measure in the target
population (17, 18). Only Kazuyoshi et al. (19)
566
made the psychometric validation of an alternate
version of the chewing function questionnaire
(CFQ), but only questions considering specific Jap-
anese food had been asked. That makes the ques-
tionnaire unfavorable for the use in other cultures.
Obviously (although the chewing is one of the
most important functions of the orofacial system),
there is a lack of a specific unidimensional instru-
ment for the patients’ self-perceived chewing func-
tion. Some questions related to psychosocial
consequences of impaired chewing function are
contained in different questionnaires including the
OHIP14 and the OHIP49 questionnaires (8, 17, 20).
Therefore, the objective of this study was to
develop an unidimensional questionnaire for the
patient’s self-assessment of chewing function and
to test its psychometric properties.
Materials and methods
The study was conducted at the School of Dental
Medicine, University of Zagreb, from May 2011 up
to May 2012.
The development of the questionnaire
The development of the questionnaire was done in
three phases: item generation, item selection, and
the reduction of the items.
As a first step, a bibliography review was con-
ducted. We searched the MEDLINE database of
the National Library of Medicine to identify the
most relevant questions and the most important
types of foods used for the assessment of a chew-
ing function, with the aim to develop a question-
naire. The search terms used were as follows:
mastication, chewing, chewing function, CFQ,
mastication assessment, chewing assessment, self-
rated mastication, self-perceived mastication, and
chewing ability. A focus group of dental profes-
sionals was formed, which included five specialists
of prosthodontics and two general dental practitio-
ners. Also, a group of 15 patients who have already
received prosthodontic treatment participated
(sample A, Table 1). They had been interviewed to
generate a pool of items that would be relevant to
describe their chewing function (ability and diffi-
culties related to chewing). A total of 30 questions
were generated.
A focus group of dental professionals discussed
about suggested items and reduced them with a
purpose to create a short, clear, easily understood
questionnaire. It was decided to use a scale format
 Chewing function questionnaire

internal consistency, test–retest reliability (n = 30)

internal consistency, test–retest reliability (n = 30)

D, F: Individuals with natural teeth without any problems at regular recall examinations at the Department of Cariesology, School of Dental Medicine, University of
from 0 to 4 (5-point Likert scale) with higher scores
indicating more severe problems. Twenty-one

Convergent validity, discriminative validity,

Convergent validity, discriminative validity,

items remained, and the pilot testing was carried
out in a group of 16 prosthodontic patients (sample
B, Table 1) and 8 dental students (sample C,
Table 1), and another four items were eliminated
after the pilot testing. A total of 17 items remained.
As the questionnaire should differentiate between
population with natural teeth (NT) and those with
removable denture wearers (RDWs), both groups
Type of investigation

were included in the study. One hundred subjects

Responsiveness
were included in the NT group (sample D, Table 1)
Factor analysis
Factor analysis
Focus group

Pilot testing

Pilot testing

and one hundred in the RDWs group (sample E,

Table 1), a total of two hundred participants.
After the questionnaire had been filled in, the
floor and ceiling effects (items with extreme means
or near-zero variances) were identified. The item
Convenience sample

Convenience sample

Convenience sample
Convenience sample
Convenience sample
Convenience sample

Convenience sample

Convenience sample

total correlation was computed. Factor analysis

Sampling method

was obtained to investigate the dimensionality of

the questionnaire. For item extraction, the principal
components extraction method and varimax rota-
tion were used. ‘Eigenvalue greater than one’ rule
determined the number of factors, and the excess
Table 1. Overview of samples (number, age, gender), sampling methods, and type of investigation

of the questions was omitted (21).

Percentage

A, B, E, G, H: Department of Prosthodontics, School of Dental Medicine, University of Zagreb.

The final version of the CFQ was unidimensional

women

with 10 items included. After development of the

CFQ, the next step was to test its psychometric
66

62

50
50
47
46

46

65

properties.
Age range

Psychometric properties
41–87

42–83

18–46
36–81
19–51

34–85

43–78
21-26

For that purpose, another 224 participants were

included. Subjects included in a psychometric vali-
Age mean

dation of the CFQ are also listed in Table 1. The

59.3 (8.7)

60.2 (4.7)

23 (2.5)
24.6 (5.7)
59.3 (4.7)
34.3 (7.2)

62.4 (9.7)

59.1 (8.2)

convergent and the discriminative validity were

(SD)

tested in 200 individuals, 100 of them were NT

group (sample F, Table 1) and another 100 were
15

16

8
100
100
100

100

24

RDWs group (sample G, Table 1). The reliability

N

C: School of Dental Medicine, University of Zagreb.

(internal consistency) of the CFQ was also tested

(H) Patients with prosthodontic treatment needs

[200 subjects, 100 of them with NT (sample F) and

100 of them RDWs (sample G)]. Another measure
of reliability (test–retest) was tested by including
(A) Patients who have already received

(B) Patients who have already received

60 participants (30 with NT from sample F and 30

RDWs from sample G) who filled in the question-
(G) Removable denture wearers
(E) Removable denture wearers

naire twice in the time interval of at least 2 weeks.

prosthodontic treatment

prosthodontic treatment

To test the responsiveness, 24 patients (sample H,

Table 1) who received a new prosthodontic treat-
ment were included.
(C) Dental students
(D) Natural teeth

(F) Natural teeth

Validity
The convergent validity was determined from the
association between the self-reported general satis-
Zagreb.
Sample

faction with chewing (assessed using the same

Likert’s scale) and the CFQ each item’s score, as

567
Persic et al.
well as with CFQ summary score by using the
Spearman’s rank correlation (22). The convergent
validity was also determined by using the sum-
mary score of the five items related to chewing
picked up from the OHIP49 questionnaire (items 1,
28, 29, 30, and 32) (5).
The discriminative validity was tested by compar-
ing the CFQ summary scores between the two
groups [RDWs (sample G) and NT groups (sample
F)]. Relevant and statistically significant differences
were expected between the two groups with lower
summary scores in the NT group.
Reliability
Two types of reliability were assessed – the test–
retest reliability and the internal consistency. The
internal consistency was assessed by calculating the
Cronbach’s alpha coefficient and the average
interitem correlation for the CFQ scores (23). The
Cronbach’s alpha coefficient and the average inter-
item correlation were calculated for all cases and
separately for cases with NT and RDWs (samples
F and G).
The test–retest reliability was assessed by calcu-
lating the intraclass correlation coefficients (ICCs)
based on the one-way repeated measures analysis
of variance (ANOVA). None of the subjects received
any treatment during the observed period and had
no problems considering teeth, dentures or soft tis-
sues in oral cavity. It was predicted that the CFQ
scores would not change during the two-week
period without any oral treatment. The ICCs were
calculated according to the Shrout and Fleiss’s
method (24).
Responsiveness
To test the responsiveness, 24 patients (sample H)
completed the CFQ twice, that is, before the treat-
ment had begun and 1 month after the prosth-
odontic treatment had been completed. A total of
12 patients received new complete dentures, five
patients received complete dentures with mandib-
ular dentures supported by four mini dental
implants, and the rest of patients (seven patients)
received partial removable dentures. The signifi-
cance of the difference in the CFQ including the
summary scores between the baseline and the fol-
low-up was tested by using the paired t-test and by
calculating the standardized effect size. The stan-
dardized effect size was calculated according to
Allen et al. as follows: mean (baseline CFQ score –
follow-up CFQ score)/standard deviation of the
baseline CFQ score (25).
568
The study was approved by the institutional eth-
ics committee. Each participant received a thor-
ough verbal explanation of the aims of the study.
Only those who provided a written informed con-
sent were included.
Statistical analysis
Statistical analysis was performed using Statistical
Package for the Social Sciences, version 17.0, for
Windows (SPSS Inc., Chicago, IL, USA) and Micro-
soft Office Excel 2003 (Microsoft, Seattle, WA,
USA). The level of significance was set at 95%
probability (P < 0.05).
Results
Chewing function questionnaire generation
and factor analysis
The initial statistical analysis (including 17 items)
revealed that the items had been grouped in three
different dimensions (components). The first com-
ponent explained 48.96% of the variance, the sec-
ond component explained 7.64%, and the third
component explained 5.91%.
By eliminating items with complexities greater
than one (items with large correlations in more
than one component) or those items with much
larger correlations in the second or the third com-
ponent than in the first component, a total of six
items were eliminated. One more item was addi-
tionally eliminated (item with the smallest corre-
lation). The final questionnaire included only 10
items that are presented in Table 2 (in Croatian
and English languages). However, in this form, it
represented the unidimensional (one-component)
questionnaire with a total of 63.67% variance
explained (Table 3). Table 4 also shows the mean
values together with standard deviations of each
item, as well as relative frequencies of the floor
and ceiling scores.
The total score of the CFQ has maximum value
of 40, which represents maximum difficulties,
while 0 score represents the absence of any prob-
lems or absence of any chewing difficulties.
Reliability
When all of the subjects from the samples F and G
were included in the analysis (n = 200), the Cron-
bach’s alpha coefficient was 0.916, and the average
interitem correlation was 0.59. For the NT (sample
F), the Cronbach’s alpha coefficient was 0.742,
and the average interitem correlation was 0.23 and
 Chewing function questionnaire

Table 2. Ten-item unidimensional chewing function (mastication) questionnaire: final version

Item
(Question) Croatian language English translation Extraction
Q1 Imate li poteskoca prilikom zvakanja jabuke/ Have you had any difficulty chewing apples / 0.89
mrkve ili hrane slicne konzistencije? raw carrots, or foods of similar consistency?
Q2 Imate li poteskoca prilikom zvakanja slanine, Have you had any difficulty chewing bacon/ 0.71
prsuta, pecenog ili pohanog mesa ili hrane smoked ham/ baked or fried firm meat,
slicne konzistencije (tvrde hrane)? or foods of similar consistency?
Q3 Imate li poteskoca prilikom zvakanja cajnih Have you had any difficulties chewing 0.47
keksa, krekera, dvopeka ili hrane slicne biscuits, crackers, tea biscuits, or foods of
konzistencije? similar consistency?
Q4 Imate li poteskoca prilikom zvakanja svjezeg Have you had any difficulty chewing fresh 0.49
kruha, krafni ili hrane slicne konzistencije? bread, doughnut or foods of similar consistency?
Q5 Imate li poteskoca prilikom zvakanja Have you had any difficulty chewing nuts / 0.84
ljesnjaka/ oraha/ badema/ makadamije/ walnuts /almonds /macadamia/peanuts,
kikirikija ili hrane slicne konzistencije? or similar food?
Q6 Imate li poteskoca prilikom zvakanja zelene Have you had any difficulty chewing lettuce, 0.63
salate, sirovog kupusa ili hrane slicne raw cabbage, or similar food?
konzistencije?
Q7 Osjecate li nesigurnost prilikom zvakanja? Have you felt insecure when you are 0.56
chewing?
Q8 Imate li poteskoca prilikom odgrizanja hrane? Have you had any difficulty when biting 0.75
different foods (food incision)?
Q9 Imate li osjecaj da vam hrana negdje zaostaje Have you noticed food catching or food 0.55
za vrijeme ili poslije zvakanja? remaining sticked between or on your teeth
or dentures during or after meals?
Q10 Imate li poteskoca prilikom zvakanja Have you had any difficulty chewing 0.51
zvakace gume? chewing gum?
Extraction method: principal component analysis.

Table 3. Total variance explained in the final 10-item chewing function questionnaire (CFQ) (Samples D and E from
Table 1)
Initial eigenvalues Extraction sums of squared loadings
Percentage Cumulative Percentage Cumulative
Component Total of variance percentage Total of variance percentage
1 6.36 63.66 63.66 6.36 63.66 63.66
2 0.85 8.52 72.19
3 0.58 5.86 78.05
4 0.57 5.70 83.75
5 0.45 4.53 88.29
6 0.38 3.83 92.12
7 0.26 2.69 94.82
8 0.23 2.29 97.11
9 0.17 1.77 98.89
10 0.11 1.10 100.00
Extraction method: principal component analysis.

for the RDWs (sample G), the Cronbach’s alpha
coefficient was 0.852, and the average interitem
correlation was 0.37.
When the test–retest reliability was evaluated, the
95% confidence intervals of the mean values were
computed. Mean differences between the first and
the second CFQ questionnaires, ICC, 95% confi-
dence intervals of the mean values, and the level of
significance are presented in Table 5. The CFQ
showed no significant difference for each item and
no significant differences for the summary scores
(P > 0.05, NS). The ICCs were satisfactory.
Validity
The convergent validity was verified by a significant
positive association (P < 0.001) between the gen-
eral single question (self-reported difficulties dur-
ing chewing) and the CFQ each item’s score, as

569
Persic et al.

Table 4. Component matrix (a) with mean values (x) Table 6. Significance of the difference between the natu-
and standard deviations (SD) and the percentages of ral teeth (NT) group (sample F, N = 100) and the remov-
floor and ceiling effects of each item able denture wearers group (RDWs, sample G, N = 100)
Ceiling Item Mean difference t
Floor effect effects
Component percentage percentage Q1 2.06 16.19*
Item 1 X (SD) of scores 0 of scores 4 Q2 1.63 12.23*
Q3 0.66 5.63*
Q1 0.93 1.24 (1.37) 45 7.5 Q4 1.02 8.66*
Q2 0.84 1.11 (1.25) 44 5.0 Q5 2.40 16.75*
Q3 0.68 0.52 (0.89) 69 0.5 Q6 1.25 9.26*
Q4 0.70 0.65 (0.98) 62 1.0 Q7 1.09 8.31*
Q5 0.91 1.57 (1.57) 38.5 20.5 Q8 2.03 13.69*
Q6 0.79 0.70 (1.14) 69 1.5 Q9 1.65 10.41*
Q7 0.74 0.71 (1.08) 62 2.0 Q10 2.54 22.12*
Q8 0.86 1.40 (1.46) 41.5 12.5 CFQ summary score 16.33 19.23*
Q9 0.74 1.72 (1.39) 27 13.5
Q10 0.71 1.39 (1.51) 47 12.5 *=P < 0.001.

Extraction method: principal component analysis. Responsiveness

(a) one component extracted.
well as with the CFQ summary score (Table 5).
Additionally, the convergent validity was also veri-
fied by a significant positive association (P < 0.001)
between the summary score of the five items
related to chewing picked up from the OHIP49
questionnaire (item number 1, 28, 29, 30, and 32)
and the CFQ each item’s and summary scores
(Table 5).
The discriminative validity was tested by com-
paring the CFQ each item’s and summary scores
between the NT group (sample F) and the RDWs
(sample G) group using Student’s independent
t-test. The results are presented in Table 6. Signi-
ficant difference was found between the NT
group and the RDWs group (P < 0.001), as
expected.
To test the responsiveness of the CFQ, the ques-
tionnaire had been filled in twice: prior the treat-
ment and 1 month after the treatment. The results
are presented in Table 7. Mean change score after
the treatment was 10.67 (SD = 8.47) (P < 0.001) for
the CFQ summary score. The standardized effect
size for the CFQ summary score was 0.95.
Discussion
We aimed to develop internationally acceptable
simple unidimensional questionnaire for the
assessment of a chewing function. After a bibliog-
raphy review and after a discussion with patients,
a pool of 30 items relevant to describe their chew-
ing function was included. It was decided to use
the five-point Likert scale (0–4) for grading, as it

Table 5. Test–retest reliability for the chewing function questionnaire (CFQ) and convergent validity: Spearman’s
coefficient of correlation
Test–retest reliability Convergent validity
OHIP summary score
CFQ (0–4) Mean 95% confidence General question (questions related
Questionnaire ICC difference interval t P (chewing difficulties) to chewing)
Q1 0.953 0.083 0.03  0.19 1.52 0.133 NS 0.65* 0.63*
Q2 0.966 0.033 0.10  0.03 1.00 0.321 NS 0.58* 0.71*
Q3 0.945 0.050 0.02  0.12 1.35 0.182 NS 0.57* 0.59*
Q4 0.980 0.000 0.05  0.05 0.00 1.000 NS 0.48* 0.51*
Q5 0.967 0.067 0.04  0.17 1.27 0.219 NS 0.60* 0.62*
Q6 0.966 0.033 0.10  0.03 1.00 0.321 NS 0.47* 0.55*
Q7 0.973 0.017 0.06  0.09 0.44 0.669 NS 0.60* 0.74*
Q8 0.952 0.067 0.05  0.18 1.16 0.252 NS 0.53* 0.60*
Q9 0.952 0.067 0.17  0.04 1.27 0.219 NS 0.50* 0.57*
Q10 0.985 0.050 0.01  0.11 1.76 0.083 NS 0.36* 0.48*
Summary score 0.991 0.200 0.55  0.15 1.14 0.260 NS 0.628* 0.721*
NS, not significant (P > 0.05); ICC, intraclass correlation coefficients.
*=P < 0.001.

570
 Chewing function questionnaire

Table 7. Responsiveness of the chewing function ques- were about feeling insecure during mastication
tionnaire (CFQ) tested on the 24 patients (sample H) (Q7), having difficulty when biting food (Q8), and
Before After about food sticking and catching on the teeth or
treatment treatment dentures (Q9). However, in this form, it repre-
Items x SD x SD t P
sented the unidimensional (one-component) ques-
tionnaire with a total of 63.67% variance explained
Q1 1.58 1.35 0.67 0.87 3.70 0.001* (Tables 3 and 4).
Q2 1.92 1.53 0.58 0.78 4.46 <0.001*
Q3 1.00 1.18 0.08 0.28 4.24 <0.001*
However, the questionnaires without appropri-
Q4 0.92 1.21 0.25 0.61 3.39 0.003* ate psychometric properties should not be used
Q5 2.08 1.47 0.58 0.88 6.43 <0.001* (17, 18, 20, 27, 28). Therefore, the next step of this
Q6 1.25 1.33 0.25 0.61 4.15 <0.001* study was to test the psychometric properties of
Q7 1.00 1.25 0.08 0.28 3.94 0.001*
the CFQ. We assessed the reliability, validity, and
Q8 1.92 1.47 0.33 0.76 6.05 <0.001*
Q9 2.58 1.64 1.42 1.14 4.61 <0.001* the responsiveness of the CFQ.
Q10 1.75 1.57 1.08 1.14 5.13 <0.001* The methods of assessing reliability of a ques-
CFQ summary 16.00 11.25 5.33 5.44 6.16 <0.000* tionnaire involve determining the extent to which
score
the test produces consistent results on retesting
*=P < 0.010; degree of freedom (d.f.) = 23. (test–retest); the relative accuracy of a test at a
given time (alternate forms); the internal consis-
had been proved to be a simple scale, yet broad tency of the items (split half); and the degree of
enough to describe precisely the differences. It has agreement between two examiners (interscorer
also been proved that the 5-point Likert scale has agreement) (29). The internal consistency examines
better test–retest properties than the broader scales whether several items that measure the same gen-
(18, 26). eral construct produce similar scores. The Cron-
After initial statistical analysis, items with com- bach’s alpha is a summary statistics, which
plexities >1 (items with large correlations in more captures the extent of agreement between all possi-
than one component) or those items with much lar- ble subsets of questions (23). This study tested the
ger correlations in the second or the third compo- reliability of the questionnaire using the Cron-
nent than in the first component were eliminated. bach’s alpha and the test–retest approach. The
Those were question numbers 1, 2, 3, 10, 13, and 14 Cronbach’s alpha values >0.80 indicate a reliable
[Q1 was the question considering simultaneous scale, although values >0.70 are also acceptable
tooth contacts on the left and the right side when (23). The Cronbach’s alpha coefficient obtained in
closing mouth; Q2 was the question considering the present study was 0.916 for all of the subjects,
pain in oral cavity during mastication; Q3 was the 0.742 for the NT group, and 0.852 for the RDWs. It
item considering eating soft food like banana or has been considered as satisfactory. Moreover, the
yogurt (it also showed floor effect, that is, zero CFQ showed also satisfactory test–retest properties
score in almost all participants); Q10 was the item and no significant difference between the two test
about the sense of the properly diminished food administrations, neither for each of the items nor
when swallowing; Q13 was the question related to for the summary score (Table 5, P > 0.05).
a sense of pain in masticatory muscles; and Q14 Another important psychometric property of a
was the item about a sense of a premature contact questionnaire is its validity. Construct validity
(or noise) when teeth come together during chew- defines the ability to seek agreement between a the-
ing]. In the further step, the item 12 was also elimi- oretical concept and a specific measuring proce-
nated (Q12 was about interrupting meals due to dure. Convergent validity investigates how closely
masticatory problems) because of the lowest corre- the new scale is related to other measures of the
lation. same construct (24). Therefore, the results of the
The final questionnaire included only ten items CFQ were compared with the test measuring simi-
that are presented in Table 2 (in Croatian and Eng- lar clinical properties (one general question consid-
lish languages). The items were related to the ering possible chewing difficulties and five
assessment of chewing difficulties when chewing questions related to chewing from the OHIP49 (5).
different common foods like apple or carrots (Q1), There was a significant positive association
meat or bacon (Q2), biscuits or crackers (Q3), fresh (P < 0.001) between the self-reported chewing
bread (Q4), different nuts (Q5), lettuce or raw function difficulties and the mean score of each
cabbage (Q6), chewing gum (Q10), or the items item of the CFQ, as well as with the mean

571
Persic et al.
summary score of the CFQ (Table 5). Moreover,
there was also a significant positive association
(P < 0.001) between the summary score of the five
questions related to chewing from the OHIP49
questionnaire and the mean score of each item of
the CFQ, as well as with the mean summary score
of the CFQ. The Spearman’s coefficient of correla-
tion varied between 0.36 and 0.74, respectively
(Table 5).
Discriminative validity is another way to measure
construct validity, and it describes the ability to
discriminate between groups with different treat-
ment needs (30). We tested the discriminative
validity between the two groups included in the
study (the NT group and the RDWs) using the Stu-
dent’s independent t-test. We predicted that the
group with NT would have significantly lower
CFQ scores than RDWs. The obtained results
showed significantly higher each item’s score and
the CFQ summary score in the RDWs (P < 0.001),
which confirmed our hypothesis (Table 6).
Responsiveness measures the response between
the two administrations of the same test caused by
a treatment procedure. We supposed that after the
prosthodontic treatment, the chewing ability
would significantly improve, that is, that the CFQ
scores would be lower. For that purpose, 24
patients requiring prosthodontic treatment were
included. Our results confirmed the satisfactory
responsiveness to appropriate treatment for the
CFQ (Table 7). All questions showed significant
improvement (lower scores) in comparison with
pretreatment results (P  0.003). According to Co-
hen, the effect size of 0.20 is considered small, 0.50
moderate, and 0.80 large (31). The effect size for
the CFQ summary score was large. It was 0.95.
The overall results for the psychometric proper-
ties of the CFQ proved the questionnaire to be sat-
isfactory and ready for wide use. The CFQ is an
unidimensional questionnaire developed to assess
the chewing function. Some other previously
developed questionnaires such as OHIP, JFLS, or
MFIQ also contain some questions considering
chewing function. For example, the OHIP ques-
tionnaire contains five items considering chewing
(5). However, those questions are not related to dif-
ficulties during chewing specific kind of food, they
are more general questions regarding chewing
function, and they also reflect individual’s social
and psychological attitude. The JFLS questionnaire,
similar to the newly developed CFQ questionnaire,
contains some questions related to chewing differ-
ent kinds of food (such as chewing crackers, meat,
572
hard food), but it also contains questions regarding
other oral functions (such as smiling, kissing, talk-
ing, frowning, laughing, etc.) and the psychosocial
impact of that functions to an individual. There-
fore, the JFLS is a multidimensional questionnaire
(2). The MFIQ, similar to the CFQ, assesses difficul-
ties during chewing food of different consistencies
(such as peanuts, almonds, raw carrot, meat), but
is also a multidimensional questionnaire evaluat-
ing mandibular functions such as speaking, kiss-
ing, yawning, laughing, and drinking (6). The CFQ
is the first unidimensional questionnaire related to
a chewing function. The limitation of this study is a
convenient sample used for item generation and
for validation. The questionnaire is convenient for
the majority of people without any diet restrictions,
that is, population with normal dietary habits.
However, it is not convenient for individuals with
special diet requirements, such as soft diet or for
vegetarians, or for people eating some specific
kinds of food.
Conclusion
The CFQ was developed by authors to assess the
direct or primary impacts of impaired chewing
function in prosthodontic patients. It was con-
structed to reflect patients’ perceived chewing
function values in order not to mix chewing func-
tion disorders with psychosocial impact of a dis-
turbed chewing function and its influence to a
patient’s quality of life. The CFQ was developed
based on patients’ opinion together with input
from dental professionals. The CFQ consists of 10
items that create the CFQ summary score ranging
from 0 to 40. The CFQ showed good psychometric
properties.
The authors recommend the use of the CFQ in
daily practice as well as in research studies, alone
or together with other questionnaires.
Acknowledgements
This study was supported by the grant 065-0650446-0420
of the Croatian Ministry of Science, Education, and
Sports.
References
1. Tortopidis D, Hatzikyriakos A, Kokoti M, Menexes
G, Tsiggos N. Evaluation of the relationship between

subjects’ perception and professional assessment of
esthetic treatment needs. J Esthet Restor Dent
2007;19:154–62.
2. Ohrbach R, Larsson P, List T. The jaw functional lim-
itation scale: development, reliability, and validity of
8-item and 20-item versions. J Orofac Pain
2008;22:219–30.
3. Larsson P, John MT, Nilner K, Bondemark L, List T.
Development of an Orofacial Esthetic Scale in prosth-
odontic patients. Int J Prosthodont 2010;23:249–56.
4. Atchison KA, Dolan TA. Development of the Geriat-
ric Oral Health Assessment Index. J Dent Educ
1990;54:680–7.
5. Slade GD, Spencer AJ. Development and the evalua-
tion of the Oral Health Impact Profile. Community
Dent Health 1994;11:3–11.
6. Stegenga B, de Bont LG, de Leeuw R, Boering G.
Assessment of mandibular function impairment
associated with temporomandibular joint osteoar-
throsis and internal derangement. J Orofac Pain
1993;7:183–95.
7. Montero J, Bravo M, L opez-Valverde A. Develop-
ment of a specific indicator of the well-being of wear-
ers of removable dentures. Community Dent Oral
Epidemiol 2011;39:515–24.
8. Brennan DS, Spencer AJ, Roberts-Thomson KF.
Tooth loss, chewing ability and quality of life. Qual
Life Res 2008;17:227–35.
9. Manly RS, Braley LC. Masticatory performance and
efficiency. J Dent Res 1950;29:448–62.
10. Hirai T, Ishijima T, Koshino H, Anzai T. Age-related
change of masticatory function in complete denture
wearers: evaluation by a sieving method with
peanuts and a food intake questionnaire method. Int
J Prosthodont 1994;7:454–60.
11. Kobayashi Y, Shiga H, Arakawa I, Yokoyama M. The
effectiveness of measuring glucose extraction for
estimating masticatory performance. Prosthodont
Res Pract 2006;5:104–8.
12. Sato H, Fueki K, Sueda S, Sato S, Shiozaki T, Kato M
et al. A new and simple method for evaluating mas-
ticatory function using newly developed artificial
test food. J Oral Rehabil 2003;30:68–73.
13. Sato Y, Minagi S, Akagawa Y, Nagasawa T. An eval-
uation of chewing function of complete denture
wearers. J Prosthet Dent 1989;62:50–3.
14. Nguyen TC, Witter DJ, Bronkhorst EM, Gerritsen
AE, Creugers NH. Chewing ability and dental func-
tional status. Int J Prosthodont 2011;24:428–36.
15. Kurita H, Ohtsuka A, Kurashina K, Kopp S. Chew-
ing ability as a parameter for evaluating the disabil-
ity of patients with temporomandibular disorders.
J Oral Rehabil 2001;28:463–5.
16. Akpata E, Otoh E, Enwonwu C, Adeleke O, Joshipu-
ra K. Tooth loss, chewing habits, and food choices
View publication stats
Chewing function questionnaire
among older Nigerians in Plateau State: a prelimin-
ary study. Community Dent Oral Epidemiol 2011;39:
409–15.
17. 
Rener-Sitar K, Petricevic N, Celebi c A, Marion L.
Psychometric Properties of Croatian and Slovenian
Short Form of Oral Health Impact Profile Question-
naires. Croat Med J 2008;49:536–44.
18. Persic S, Milardovic S, Mehulic K, Celebic A. Psycho-
metric properties of the Croatian version of the
Orofacial Esthetic Scale and suggestions for modifi-
cation. Int J Prosthodont 2011;24:523–33.
19. Kazuyoshi B, John MT, Inukai M, Aridome K, Iga-
rahsi Y. Validating an alternate version of the chew-
ing function questionnaire in partially dentate
patients. BMC Oral Health 2009;16:9–9.
20. 
Rener-Sitar K, Celebi c A, Petricevic N, Papic M, Sap-
undzhiev D, Kansky A et al. The Slovenian version
of the Oral Health Impact Profile Questionnaire
(OHIP-SVN): translation and psychometric proper-
ties. Coll Antropol 2009;33:1177–83.
21. Yeomans KA, Golder PA. The Guttman-Kaiser crite-
rion as a predictor of the number of common factors.
The Statistician 1982;31:221–9.
22. Cohen J. Statistical Power Analysis for the Behav-
ioral Sciences. Hillsdale, NY: Lawrence Erlbaum;
1988.
23. Bland JM, Altman DG. Statistical methods for assess-
ing agreement between two methods of clinical mea-
surement. Lancet 1986;1:307–10.
24. Fleiss J. The Design and Analysis of Clinical Experi-
ments. New York: Wiley & Sons; 1986.
25. Allen PF, McMillan AS, Locker D. An assessment of
sensitivity to change of the Oral Health Impact Pro-
file in a clinical trial. Community Dent Oral Epidemi-
ol 2001;29:175–82.
26. Lissitz RW, Green SB. Effect of the number of scale
points on reliability: a Monte Carlo approach. J Appl
Psychol 1975;60:10–3.
27. 
Petricevic N, Celebi c A, Rener-Sitar K. A 3-year lon-
gitudinal study of quality-of-life outcomes of elderly
patients with implant- and tooth-supported fixed
partial dentures in posterior dental regions. Gerod-
ontology 2012;29:e956–63.
28. 
Petricevic N, Celebi c A, Papic M, Rener-Sitar K. The
Croatian version of the Oral Health Impact Profile
Questionnaire. Coll Antropol 2009;33:841–7.
29. Groth-Marnat G. Handbook of psychological assess-
ment, 4th edn. Hoboken, New Jersey: Wiley & Sons;
2003.
30. Streiner D, Norman G. Health Measurement Scales-
A Practical Guide to Their Development and Use.
Oxford: Oxford University Press; 2008.
31. Cohen J. Statistical Power Analysis for the Behav-
ioral Sciences. Hillsdale, NY: Lawrence Erlbaum;
1988.
573