Online quality and

readability assessment
of Early childhood caries
P. E. Ayala Aguirre1,2, I. Aníbal1,
A. P. Strieder1, M. Lotto1,
V. Lopes Rizzato1,

information available on A. F. Pereira Cruvinel3, D. Rios1,

T. Cruvinel1

websites from distinct 1

Department of Pediatric Dentistry, Orthodontics
and Public Health, Bauru School of Dentistry,

countries:
University of São Paulo, Bauru, Brazil
2
Universidad de Especialidades Espíritu Santo
(UEES), Samborondón, Ecuador

A cross-sectional study 3
Discipline of Public Health, School of Medicine,
Federal University of Fronteira Sul, Chapecó,
Brazil

DOI 10.23804/ejpd.2022.23.01.03 E-mail: [email protected]

Abstract cavitated lesions), missing (due to caries), or filled tooth

surfaces in any primary tooth in a child under the age of six”
[Pitts et al., 2019]. It impacts children’s and families’ quality
Aim To assess the quality and readability of ECC-related Web of life due to its negative consequences, such as chronic pain,
information available in English, Spanish, and Brazilian Portuguese
infections, sleeping disorders, loss of school days, and tooth
language.
extraction [Pitts et al., 2019; GBD Oral Disorders Collaborators
Methods This study assessed the quality and readability of
information related to ECC in three different languages found on et al., 2020]. In theory, these specific problems and
the most popular Internet search engines worldwide. Websites circumstances lead parents and caregivers to recognise the
were retrieved from different search engines using specific need for a particular information, e.g., recommendations for
strategies. DISCERN questionnaire, JAMA benchmark criteria, and the management of ECC [Yardi et al., 2018].
language-based readability formulas were used by two independent Although physicians are still the number one source of
investigators to evaluate the quality and readability of websites. health information [Ipsos, 2018], the sense of self-efficacy
Also, contents were categorised according to aetiology, prevention, in relation to selfcare [Bandura, 1988], next to the Internet’s
and treatment of ECC. The statistical analysis was performed using pervasiveness, could trigger parents and caregivers’ seeking
Spearman’s rank correlation coefficient, hierarchical clustering behaviour for online oral health consultations as means to
analysis by Ward’s minimum variance method, and Mann-Whitney fulfill their needs for information by themselves. Nevertheless,
U test. P values < 0.05 were considered significant. the access to digital health-related mis- or dis-information
Results Digital contents of 177 websites were considered of easy could hamper the primary prevention of ECC, because of
readability and low quality, without differences between websites
their respective unintentional or deliberated incompleteness,
produced by health- or non-health-related authors. Also, websites
with ≥ 3 categories of information, or those that recommended
uncertainty, or ambiguity natures [Stahl, 2006], in combination
dental visits or fluoride toothpastes presented higher DISCERN with the challenges represented for socially marginalised
scores than their counterparts. young parents in assessing the quality of information
Conclusion ECC-related digital contents were considered simple, [Greyson, 2018]. Additionally, low-quality information could
accessible and of poor quality, independently of their language and negatively impact on the person-professional relationship,
authorship. These findings indicate the importance of professional especially among individuals who ignore the expertise of
counseling to empower parents in selecting and consuming adequate dentists for conflicting with their own beliefs confirmed with
information towards the improvement of children’s oral health what was found online [Tan and Goonawardene, 2017; Glick,
2017].
Based on the aforementioned reasons, the aim of this study
was to analyse the quality and readability of ECC-related
KEYWORDS Early Childhood Caries; Prevention; Health behaviour; information available on the Internet.
Quality of information; Internet; eHealth.

Methods

Design
Introduction The present study assessed the quality and readability of
information related to ECC in three different languages
Early childhood caries (ECC) is a chronic disease defined (English, Spanish and Brazilian Portuguese) found on the most
as “the presence of one or more decayed (non-cavitated or popular Internet search engines worldwide. Two instruments

European Journal of Paediatric Dentistry vol. 23/1-2022 15

AYALA AGUIRRE P. E. ET AL.

were used to analyse the quality of information, the DISCERN for each criterion fulfilled, with a total score of 0 to 4) [Silberg
questionnaire and the Journal of American Medical Association et al., 1997].
(JAMA) benchmark criteria. For the readability analysis, specific The DISCERN instrument is a standardised questionnaire
metrics were used for each language. Websites were consisting of 16 Likert-scale questions, where 1 represents
categorised regarding to their content and dichotomised not meeting the criteria and 5 completely meets the criteria,
according to the nature of their authorship, recommendation divided into three sections: the first section describes the
of fluoride toothpaste usage and dental visits. reliability of the content (questions 1 to 8), the second section
Since federal regulations consider that research using analyses the information regarded treatment options
publicly available data does not involve human subjects, this (questions 9 to 15), and the third section represents an overall
protocol did not require institutional review board approval assessment of the website (question 16). Only the results of
from the Council of Ethics in Human Research of the Bauru the first and second sections are commonly used to determine
School of Dentistry, University of São Paulo (Brazil). The the quality of information of written materials, categorised
present study followed the STrengthening the Reporting of in very poor (scores between 15 and 26), poor (27 to 38), fair
OBservational Studies in Epidemiology (STROBE) checklist for (39 to 50), good (51 to 62), and excellent (63 to 75) [Charnock
cross-sectional studies. et al., 1999].
Two examiners independently assessed all websites (PEAA
Search strategy and data collection and IA). Regarding both instruments, when a divergent
First, all cookies and history browser were cleared of the judgment was observed between the examiners, the website
computer used for research prior data collection. Then, was re-assessed to the achievement of a consensus score.
country and language were selected on the browser’s
advanced section for each language studied: United States Readability
for English, Mexico for Spanish, and Brazil for Portuguese; The readability measures indicate the reading difficulty of
otherwise, results could be influenced by the automated set a text based on specific metrics, such as average sentence
up region. Three Internet search engines were used to retrieve length (ASL) and average number of syllables per word (ASW).
the websites according to the market share by country Those metrics were collected from the online tool Readable.
[Statcounter, 2019; Statcounter, 2019b; Statcounter, 2019a]. io (Readable.io, Bolney, England) through the Uniform
A search strategy was constructed by the association of the Resource Locator (URL) of the website or the direct input of
most commonly used terms related to ECC. For that, a list of the texts in the platform.
terms in each language was tested in the Google Trends tool Specific metrics for each language were used in this study.
to convey those with the most volume of interest. Then, The Flesch Reading Ease (FRE) and the Flesch-Kincaid Grade
several strategies associating different terms were consulted Level (FKGL) are common metrics used to evaluate the grade
until the development of a final strategy based on the higher of difficulty of English texts [Flesch, 1948; Kincaid et al., 1975].
number of results retrieved. Final data collection was The formulas predict the level of education that a person
performed on February 7th, 2019 using the strategies [“infant requires to understand a specific text: FRE = 206.835 – (1.015
tooth decay” + “nursing bottle tooth decay”] for retrieving x ASL) – (84.6 x ASW). A text scored between 90 and 100 is
English websites, [“caries de bebe”] for retrieving Spanish considered easily understandable by an average 5th grader.
websites, and [“carie em bebê” + “cárie de mamadeira” + Scores between 60 and 70 indicate a text easily understandable
“carie de mamadeira”] for retrieving Brazilian Portuguese by 8th and 9th graders, while scores between 0 and 30 are
websites. Websites with access difficulties (link errors), related to difficult texts, understandable only by graduates.
academic information (scientific papers), video links, journal The FKGL scores are equivalent to the US grade level of
news, duplicates, and nonspecific sources (Q& As, education that the reader needs to be able to comprehend
propagandas, pictures, etc.) were excluded. Before the written contents: FKGL = 0.39 (average number of words per
assessment, all links were registered on archive.today (http:// sentence) + 11.8 (average number of syllables per word) -
archivetoday.com/), an online archiving system that allows 15.59. The Fernández-Huerta Readability Formula (FHRF) is
web material to remain unaltered for future evaluation. a modified version of the Flesch Reading Ease formula for
Spanish texts in which 0 is the score for the greatest difficulty
Categorisation of websites and 100 is the score for the easiest reading, similarly to FRE
All websites were dichotomised according to the nature of (Fernandez Huerta, 1959). Its formula is expressed as FHRF =
the authorship in health- (dental professionals, clinics, 206.84 – (0.6 x number of syllables per 100 words) – (1.02 x
universities and health companies) or non-health-related number of sentences per 100 words). Brazilian Portuguese
authors (blogs and others), the diversity of contents regarding contents were evaluated by the Brazilian Portuguese version
aetiology, prevention, and treatment of ECC (higher = of the Flesch Reading Ease (FRE-BP): FRE-BP = 248.835−
detection of 3 issues; lower = detection of 2 or less issues), (84.6×syllables per word) − (1.015×words per sentence). The
the recommendation of fluoride toothpaste usage (yes/no), reading difficulty of written materials is indicated by the scores
and the recommendation of dental visits (yes/no). in very easy (75-100), easy (50-75), difficult (25-50), and very
difficult (0-25) [Martins et al., 1996].
Quality assessment
Two instruments were used for the quality assessment. The Data analysis
JAMA benchmark consists of the identification of four All data were analysed with the Statistical Package for Social
concepts on the website: authorship (author’s name, Science (v 25.0; SPSS, Chicago, USA). The internal consistency
affiliations, and credentials), attribution (effective references of DISCERN was determined by Cronbach alpha. The
of content), currentness (presence of dates of posts and interexaminer reliability was determined by the intraclass
updates of information), and disclosure (the statement of any correlation coefficient (ICC) for absolute concordance to
potential conflicts of interest) of websites, granting 1 point DISCERN and JAMA scores. The correlations between distinct

16 European Journal of Paediatric Dentistry vol. 23/1-2022

 PREVENTION AND LIFESTYLE

measures were demonstrated by the Spearman rank Mexico Brazil

USA
correlation coefficients. The websites of distinct dichotomised (English) (Spanish) (Portuguese)
natures were compared by Mann-Whitney U test. The quality

Data
Google = 88 Google = 96 Google = 97
measures of different countries were compared by Kruskal Bing = 6 Bing = 3 Bing = 1
Wallis and Dunn’s post-hoc tests. P values < 0.05 were Yahoo! = 4 Yahoo! = 1 Yahoo! = 1

considered significant for all analyses.

Screened
English Spanish Portuguese
Results
n = 98 n = 100 n = 99

A total of 297 websites were collected among the three

languages, excluding 120 links due to link errors (n = 35), English Spanish Portuguese
academic source (n = 23), video links (n = 6), news (n = 19), Link error = 12 Link error = 14 Link error = 9
Academic = 5 Academic = 6 Academic = 12
duplicates (n = 10) and nonspecific sources (n = 27). Finally, Video links = 0 Video links = 4 Video links = 2
News = 1 News = 17 News = 1
177 links met the criteria for evaluation (English n = 64, Duplicates = 8 Duplicates = 2 Duplicates = 0

Excluded
Nonspecific = 8 Nonspecific = 16 Nonspecific = 3
Spanish n = 41, and Brazilian Portuguese n = 72). Details of
exclusion criteria by language are shown on Figure 1.
DISCERN and JAMA benchmark presented good internal English Spanish Portuguese
consistencies (Cronbach’s alpha= 0.729 and 0.677, Google = 26 Google = 57 Google = 27
Bing = 4 Bing = 1 Bing = 0
respectively). The inter-examiner reliability was excellent for Yahoo! = 4 Yahoo! = 1 Yahoo! = 0
both instruments (DISCERN ICC=0.81, 95% CI: 0.72-0.87; n = 34 n = 59 n = 27
JAMA benchmark ICC=0.80, 95% CI: 0.74-0.85).
In an overall analysis, the mean of DISCERN indicated a very
poor quality of information for the three languages English Spanish Portuguese
(24.79±4.66) in concordance with the JAMA benchmark Google = 62 Google = 39 Google = 70
Evaluated

Bing = 2 Bing = 2 Bing = 1

(1.38±0.96). In total, 75.10% of the websites were classified Yahoo! = 0 Yahoo! = 0 Yahoo! = 1
n = 64 n = 41 n = 72
as very poor quality of information, 22.60% were classified
as poor quality of information and only 2.30% as fair quality
of information. There were no websites classified as either
n = 177
good or excellent quality of information. Also, the sum of
partial DISCERN scores (sections 1 and 2) were fair correlated
with the scores of section 3 (ρ=0.55, P<0.001). Regarding to
JAMA criteria, only one website published in Portuguese FIG. 1 Flowchart depicting the systematic selection of ECC-related
presented the maximum score (JAMA = 4). Ninety-four websites by language.
websites (53.11%) showed information related to the
authorship, affiliations or credentials, 143 websites (80.79%)
did not reference the content, 97 websites (54.80%) dated their counterparts. On the other hand, the authorship of
the information posted, and only 19 websites (10.73%) stated websites did not significantly influence the quality of
any potential conflict of interest. DISCERN and JAMA information (Table 2).
benchmark were weakly positive correlated (ρ=0.25, P<0.001). Most websites referred to etiology (94.40%) and prevention
Table 1 depicts the descriptive statistics of quality and of ECC (96.60%), and only 17.50% described treatment
readability scores according to distinct languages. The mean options for the disease. DISCERN scores were significantly
of FRE (63.65±7.44) and FKGL (7.85±1.61) for English websites higher in websites of higher diversity contents in comparison
indicated accessible and plain texts, easily understood by 7th with their counterparts, independently of country/language
– 8th graders. Only 2 websites were measured as difficult to (Table 2).
read. The mean of Fernandez-Huerta formula (67.37±11)
showed that Spanish websites were easy to read. Only 5
websites were considered difficult and 2 were considered Discussion and conclusion
very easy to read. Finally, the FRE-BP demonstrated that
Brazilian Portuguese websites were accessible and easy to To the best of our knowledge, this is the first study that
understand (58.83±8.47). Overall, quality and readability evaluated the quality of ECC-related information in different
scores were not correlated significantly, except to Fernández- languages. Our findings showed a predominance of accessible,
Huerta that showed weak negative correlations with DISCERN understandable, and low-quality content on the Internet,
(S1 + S2) (ρ = -0.430, P<0.001) and JAMA scores (ρ = -0.402, independently of the country and language, although the
P<0.001). quality of information and their readability were only slightly
A total of 49.20% websites were related to health authors, negatively correlated in Spanish websites. Interestingly, the
24.90% recommended fluoride toothpaste usage, and nature of the authorship did not influence the results, i.e.,
71.80% indicated dental visits. Regarding specific languages, the quality of information produced even by health-related
most websites published in Spanish (68.30%) and Brazilian authors was also very poor. On the other hand, contents with
Portuguese (59.70%) were developed by non-health authors, higher diversity presented a better-quality level of information
whereas 70.30% of websites published in English were in all languages. In general, websites recommending the use
developed by health-related authors. The DISCERN scores of fluoride dentifrices and dental visits were better qualified
were significantly higher among websites that recommended than their counterparts. These results presented a good
fluoride toothpaste usage and dental visits in comparison to interexaminer reliability and internal consistency as proven

European Journal of Paediatric Dentistry vol. 23/1-2022 17

AYALA AGUIRRE P. E. ET AL.

DISCERN
Outcomes S1a S2a S3a JAMAb FKGLc FREd FHe FRE-BPf
(S1+S2)a
Mean (SD) 15.94 (2.98) 8.84 (2.80) 1.93 (0.61) 24.79 (4.66) 1.38 (0.96) - - - -
Median 16 8 2 24 1 - - - -
Overall
Minimum 9 3 1 16 0 - - - -
Maximum 27 21 4 44 4 - - - -
Mean (SD) 16.05 (2.94) 9.25 (3.03) 2.59 (0.61) 27.13 (5.63) 1.09 (1.05) 7.85 (1.61) 63.65 (7.44) - -
USA Median 15.50 8.50 3 26 1 7.8 65.45 - -
(English) Minimum 12 7 2 20 0 5 38 - -
Maximum 27 21 4 47 3 12 78 - -
Mean (SD) 15.61 (2.87) 9.22 (3.59) 2.41 (0.63) 24.83 (5.05) 1.46 (0.87) - - 67.37 (11) -
Mexico Median 15 8 2 24 2 - - 67.29 -
(Spanish) Minimum 9 3 1 16 0 - - 43.34 -
Maximum 23 21 4 40 3 - - 94.98 -
Mean (SD) 16.04 (3.11) 8.26 (1.86) 1.97 (0.53) 24.31 (3.80) 1.58 (0.88) - - - 58.83 (8.47)
Brazil Median 16 7 2 23.50 2 - - - 59.32
(Portuguese) Minimum 9 7 1 18 0 - - - 33.24
Maximum 27 15 3 36 4 - - - 75.20
aS1, S2, and S3: 3 different sections of DISCERN.
bJAMA: The Journal of American Medical Association.
cFKGL: Flesch Kincaid Grade Level
dFRE: Flesch Reading Ease
eFH: Fernandez Huerta readability metric
fFRE-BP: Flesch Reading Ease adapted to Brazilian Portuguese.
Significant statistical differences between the groups (Mann-Whitney U test, P<.05).

TABLE 1 Descriptive statistics of scores of DISCERN, the Journal of American Medical Association benchmark, Flesch Reading Ease adapted to
Brazilian Portuguese, Fernandez Huerta, Flesch Kincaid Grade Level and Flesch Kincaid Reading Ease by language.

with a previous study [Aguirre et al., 2017]. dentistry. In addition, the better qualification of more diverse
The almost ubiquitous availability of the Internet permitted information emphasizes the importance of previous knowledge
the easy access to information, defeating spatial and temporal and professional planning to achieve contents of adequate
barriers. Not surprisingly, online search engines represent the educational performance [Kulasegarah et al., 2018].
second most used source for health information among adults Although dental caries is the most prevalent disease in the
worldwide [Ipsos, 2018]. Parents and caregivers who refer to world (GBD Oral Disorders Collaborators et al., 2020), its
websites for information regarding their children’s health impact in children could still be disregarded. For instance, oral
could rely on the content they found [Kubb and Foran, 2020], health education during pregnancy [Paglia, 2019], the need
especially if it confirms something they already believe, read of early dental visits, the importance of the maintenance of
or heard before [Meppelink et al., 2019]. Hence, the lack of primary teeth for adequate oral and general health conditions
quality control of Web information could jeopardise user’s are probably overlooked by parents [Colombo et al., 2019].
health condition based on poor-quality contents [Kubb and In this scenario, most websites described information related
Foran, 2020]. In this sense, a challenge sets when the internet to natural breastfeeding, formulas and dental hygiene as
is filled with blogs consisting in experiences, health beliefs formal “protocols” or steps that should be followed by
and opinions regarding ECC exposing their readers to parents. Hypothetically, stric t non-individualised
misinformation, e.g. “…avoid fluoride toothpastes because recommendations could dissuade parents and caregivers
they are toxic”. towards good oral health habits for not fitting in their own
Since the production of good quality information is a routines and lifestyles [Ferrazzano et al., 2019; Lotto et al.,
complex and arduous task, which involves the need of skilled 2020].
specialists, trained to attain target audiences and produce The consumption of digital contents empowers people to
accessible contents according to specific guidelines and discuss diverse aspects of their own health during the shared
instruments, the poor quality of information authored by decision-making, causing possible negative effects on the
health professionals might indicate their technical incapacity relationship with professionals when based on beliefs and
combined with a primary interest in producing oral health misinformation [Kubb and Foran, 2020]. In this scenario, when
information for advertisements and commercial profit, linked health professionals are welcoming towards their patients’
to a possible marketing sense of obligation in developing a Internet retrievals, patients are more positive about their
website for describing treatments to their patients [Mourad relationship, otherwise they seemed challenged facing a
et al., 2020]. It can be confirmed by a great volume of content negative perception, especially if they have concerns about
replication found among diverse websites. Indeed, the quality of the information gathered [Tan and
professionals seemed to be unconscious about their Goonawardene, 2017]. Acknowledging the vast volume of
responsibility in producing good quality educational oral health misinformation online and, moreover, the
information to stimulate healthy behaviours; instead, they are challenges that represent for parents to appraise good quality
more concerned to show their current practice describing of information [Greyson, 2018], health professionals have the
procedures not necessarily committed to the evidence-based responsibility to guide their patients toward clear and accurate

18 European Journal of Paediatric Dentistry vol. 23/1-2022

 PREVENTION AND LIFESTYLE

S1* S2* S3* DISCERN(S1+S2) JAMA† FKGL‡ FRE§ FH¶ FRE-BP#

Health 15.74 (2.59)a 8.87 (2.22)a 1.90 (.61)a 24.61 (3.89)a 1.33 (.97)a - - - -
Authorship
Non-health 16.14 (3.32)a 8.81 (3.28)a 1.96 (.62)a 24.96 (5.32)a 1.42 (0.96)a - - - -
Higher 16.93 (3.78)a 12.37 (3.99)a 2.44 (0.64)a 29.30 (6.41)a 1.56 (1.01)a
Overall

Diversity of contents
Lower 15.77 (2.79)a 8.21 (1.96)b 1.83 (0.56)b 23.97 (3.76)b 1.35 (0.96)a
Yes 16.27 (3.17)a 9.17 (3.01)a 1.96 (0.62)a 25.44 (5.05)a 1.43 (0.93)a - - - -
Dental visits
No 15.12 (2.27)b 8.00 (1.97 b 1.84 (0.58)a 23.12 (2.93)b 1.24 (1.04)a - - - -
Yes 17.39 (3.16)a 9.34 (3.16)a 2.14 (0.63)a 26.73 (4.62)a 1.70 (0.93)a - - - -
Fluoride toothpaste usage
No 15.47 (2.77)b 8.68 (2.67)b 1.86 (0.59)b 24.14 (4.51)b 1.27 (0.95)b - - - -
Health 15.36 (2.22)a 8.73 (2.05)a 1.80 (0.59)a 24.09 (3.58)a 1.02 (0.89)a 7.91 (1.56)a 64.11 (5.50)a - -
Authorship
Non-health 17.68 (3.77)b 10.47 (4.45)a 2.21(0.92)a 28.16 (7.32)b 1.26 (1.37)a 7.72 (1.77)a 62.57 (10.74)a - -
Higher 17.33 (4.47)a 12.56 (4.39)a 2.67 (0.71)a 29.89 (8.19)a 1.56 (1.13)a 8.24 (1.54)a 58.39 (11.08)a - -
USA (English)

Diversity of contents
Lower 15.84 (2.61)a 8.71 (2.41)b 1.80 (0.65)b 24.55 (4.29)b 1.02 (1.03)a 7.79 (1.63)a 64.51 (6.40)a - -
Yes 16.52 (3.07)a 9.58 (3.25)a 2.00 (0.72)a 26.10 (5.61)a 1.19 (1.02)a 7.75 (1.72)a 64.33 (7.40)a - -
Dental visits
No 14.62 (2.00)b 8.25 (2.05)b 1.69 (0.70)a 22.88 (3.10)b 0.81 (1.10)a 8.18 (1.22)a 61.63 (7.43)a - -
Yes 18.62 (2.87)a 10.54 (3.46)a 2.46 (0.66)a 29.15 (5.03)a 1.77 (1.09)a 7.20 (1.73)a 65.22 (9.86)a - -
Fluoride toothpaste usage
No 15.39 (2.60)b 8.92 (2.87)b 1.78 (0.67)b 24.31 (4.91)b 0.92 (0.98)b 8.02 (1.56)a 63.25 (6.75)a - -
Health 15.31 (2.75)a 9.92 (2.96)a 1.92 (0.64)a 25.23 (4.59)a 1.54 (0.88)a - - 69.20 (12.25)a -
Authorship
Non-health 15.75 (2.96)a 8.89 (3.85)a 1.82 (0.55)a 24.64 (5.32)a 1.43 (0.88)a - - 66.52 (10.49)a -
Mexico (Spanish)

Higher 15.80 (1.92)a 15.60 (5.55)a 2.60 (0.55)a 31.40 (6.91)a 1.20 (0.84)a - - 68.65 (10.23)a -
Diversity of contents
Lower 15.58 (3.00)a 8.33 (2.14)b 1.75 (0.50)b 23.92 (4.08)b 1.50 (0.88)a - - 58.13 (13.14)a -
Yes 15.81 (2.82)a 9.48 (3.79)a 1.81 (0.60) a 25.29 (5.31)a 1.55 (0.81)a - - 64.65 (9.55)a -
Dental visits
No 15.00 (3.09)a 8.40 (2.88)a 2.00 (0.47) a 23.40 (4.03)a 1.20 (1.03)a - - 75.78 (11.36)b -
Yes 17.25 (1.82)a 9.58 (4.36)a 2.00 (0.60)a 26.83 (4.30)a 1.33 (0.89)a - - 64.04 (8.30)a -
Fluoride toothpaste usage
No 14.93 (2.98)b 9.07 (3.29)a 1.79 (0.56)a 24.00 (5.17)b 1.52 (0.87)a - - 68.75 (11.79)a -
Health 16.52 (2.94)a 8.62 (2.04)a 2.03 (0.63)a 25.14 (4.04)a 1.72 (0.19)a - - - 58.04 (8.61)a
Authorship
Non-health 15.72 (3.21)a 8.02 (1.71)a 1.93 (0.46)a 23.74 (3.56)a 1.49 (0.12)a - - - 59.37 (8.43)a
razil (Portuguese)

Higher 17.08 (3.97)a 11.00 (2.24)a 2.23 (0.60)a 28.08 (4.96)a 1.69 (1.03)a - - - 56.55 (8.87)a
Diversity of contents
Lower 15.81 (2.87)a 7.66 (1.08)b 1.92 (0.50)a 23.47 (2.94)b 1.56 (0.86)a - - - 59.34 (8.37)a
Yes 16.31 (3.50)a 8.56 (1.99)a 2.02 (0.53)a 24.87 (4.25)a 1.60 (0.87)a - - - 58.95 (8.11)a
Dental visits
No 15.50 (2.06)a 7.67 (1.44)b 1.88 (0.54)a 23.17 (2.35)a 1.54 (0.93)a - - - 58.62 (9.31)a
Yes 16.63 (3.83)a 8.37 (1.50)a 2.00 (0.58)a 25.00 (3.93)a 1.89 (0.81)a - - - 56.44 (1.55)a
B

Fluoride toothpaste usage

No 15.83 (2.81)a 8.23 (1.99)a 1.96 (0.52)a 24.06 (3.75)a 1.47 (0.89)a - - - 59.69 (1.22)a
*S1, S2, and S3: 3 different sections of DISCERN.
†The Journal of American Medical Association.
‡Flesch Kincaid Grade Level.
§Flesch Reading Ease.
¶Fernandez-Huerta.
#Flesch Reading Ease adapted to Brazilian Portuguese.
Figure legend

TABLE 2 The comparison of means (SD) of quality and readability scores between dichotomised categories of websites published in distinct
languages. Different superscript letters represent significant statistical differences between dichotomised categories regarding the same
quality and readability measure (Mann-Whitney U test, P<0.05).

information, encouraging them to discuss their findings, first posting and updates shown to avoid possible hazards
especially when conditions such as ECC are behaviour for health information consumers when reading information
dependent [Pitts et al., 2019]. To improve health literacy levels, out of date (National Library of Medicine-NIH, 2018).
the greater and frequent involvement with health information The present study has some limitations. First, although the
should be stimulated, enhancing their proper capability in total score could have been influenced by the section 2 of
dealing with contents. Therefore, online search techniques the DISCERN questionnaire, when a website did not provide
as clicking restraint and lateral reading should be instructed information regarding treatment options, the independent
to support consumers in the analysis of the veracity of analysis of sections 1 and 3 also corroborated with the same
information [Wineburg and McGrew, 2019]. It consists in classification of poor quality of information. Second, although
analyzing thoroughly the links depicted before selecting a other populations are also native speakers of these three
website, scanning a webpage without diving completely in languages, the core of quality of ECC-related information
its content and opening new tabs laterally regarding the same presented here must be regarded specifically by selected
subject to compare results. Also, triangulation is a method countries, since other websites could be available for distinct
used by researchers to cross-check information consulting geographical areas. Third, both assessment instruments were
different perspectives and sources [Barnes and Vidgen, 2006], developed to assess the quality of only textual contents, i.e.,
including the use of escalation of authorities to verify any graphic elements, figures and pictures were disregarded in
information, e.g. hearing information about parenthood, this evaluation. For the same reason, 7.72% links related to
consulting first with the Internet, discussing it within their scientific papers were also excluded from this analysis;
inner circle, to finally verify it with their health care provider however, it is presumed that lay readers would have a limited
[Greyson, 2018]. Altogether, these techniques emphasise the interested in formal contents, especially due to their difficult
need to consult different sources to confirm an information in understanding technical language and terms.
of interest. In addition, the production of good quality In conclusion, the information related to ECC published in
educational contents, although a challenging task, is critical websites of Brazil, Mexico, and USA were considered of poor
to tackle misinformation. Important aspects must be quality, presenting accessible and simple contents for most
considered before the production of health content, e.g., the population. The awareness on what parents and caregivers
disclosure of objectives, authorship, references that could be are exposed to when searching for ECC online could be a
listed under de main text [Winker et al., 2000], and dates of relevant approach to improve the management of the disease,

European Journal of Paediatric Dentistry vol. 23/1-2022 19

AYALA AGUIRRE P. E. ET AL.

documents/2018-07/global-views-healthcare-report-2018-07-24.pdf
considering that the information influences the parental [Accessed May 2020].
decision-making process decisively. Therefore, paediatric › Kincaid JP, Fishburne Jr RP, Rogers RL, Chissom BS. Derivation of new
dentists should receive sufficient technical formation and readability formulas (automated readability index, fog count and flesch
stimulation to produce Web contents directly to parents and reading ease formula) for navy enlisted personnel. Institute for Simulation
children, focused on the spread of information related to and Training 1975.
good oral health habits for the prevention and control of › Kubb C, Foran HM. Online Health Information Seeking by Parents for Their
Children: Systematic Review and Agenda for Further Research. J Med
ECC. Internet Res 2020; 22, e19985.
› Kulasegarah J, McGregor K, Mahadevan M. Quality of information on
Acknowledgements the Internet—has a decade made a difference? Ir J Med Sci 2018; 187,
This research was funded by the São Paulo Research 873-876.
Foundation (#2018/02563-6). The funder had no role in the › Lotto M, Strieder AP, Ayala Aguirre PE, Andrade Moreira Machado MA,
Rios D, Cruvinel A, Cruvinel T. Parental perspectives on early childhood
study design, data collection and analysis, decision to publish, caries: a qualitative study. Int J Paediatr Dent 2020; 30(4):451-8.
or preparation of the manuscript. › Martins TB, Ghiraldelo CM, Nunes MdGV, Oliveira Junior ONd. Readability
formulas applied to textbooks in brazilian portuguese. Icmsc-Usp 1996.
Conceived and designed the study: PEAA, AFPC, TC. › Meppelink CS, Smit EG, Fransen ML, Diviani N. “I was right about
Acquired the data: PEAA, IA, TC. Analyzed and interpreted vaccination”: Confirmation bias and health literacy in online health
the data: PEAA, IA, TC. Drafted and critically revised the information seeking. J Health Commun 2019; 24, 129-140.
› Mourad M, Brendebach J, Alkilzy M, Splieth C. Choosing a paediatric
manuscript: PEEA, IA, APS, ML, VLR, AFPC, DR, TC. dentist: Factors influencing parents’ decision. Eur J Paediatr Dent 2020;
21, 74-79.
› National Library of Medicine (NIH). 2018. Evaluating Internet Health
References Information: A Tutorial [Online]. Available: https://medlineplus.gov/
webeval/webeval.html [Accessed May 2020].
› Aguirre PEA, Coelho MM, Rios D, Machado MAAM, Cruvinel AFP, › Paglia L. Oral prevention starts with the mother. Eur J Paediatr Dent 2019;
Cruvinel T. Evaluating the dental caries-related information on Brazilian 20, 173-173.
websites: qualitative study. J Med Internet Res 2017; 19, e415. › Pitts NB, Baez RJ, Diaz-Guillory C, Donly KJ, Alberto FC, McGrath C,
› Bandura A. Organisational Applications of Social Cognitive Theory. Aust J Phantumvanit P, Seow WK, Sharkov N, Songpaisan Y. Early Childhood
Manag 1988; 13, 275-302. Caries: IAPD Bangkok Declaration. J Dent Child (Chic) 2019; 86, 72.
› Barnes SJ, Vidgen RT. Data triangulation and web quality metrics: A case › Silberg WM, Lundberg GD, Musacchio RA. Assessing, controlling, and
study in e-government. Inf Manag 2006; 43, 767-777. assuring the quality of medical information on the Internet: Caveant
› Charnock D, Shepperd S, Needham G, Gann R. DISCERN: an instrument lector et viewor—Let the reader and viewer beware. JAMA 1997; 277,
for judging the quality of written consumer health information on 1244-1245.
treatment choices. J Epidemiol Community Health 1999; 53, 105-111. › Stahl BC. On the Difference or Equality of Information, Misinformation,
› Colombo S, Gallus S, Beretta M, Lugo A, Scaglioni S, Colombo P, Paglia and Disinformation: A Critical Research Perspective. Inf Sci 2006; 9.
M, Gatto R, Marzo G, Caruso S. Prevalence and determinants of early › Statcounter. 2019. Search Engine Market Share Brazil [Online]. Available:
childhood caries in Italy. Eur J Paediatr Dent 2019; 20, 267-273. http://gs.statcounter.com/search-engine-market-share/all/brazil [Accessed
› Fernandez Huerta J. Simple readability measures. Consigna 1959; 214, February 2019].
29-32. › Statcounter. 2019 -a. Search Engine Market Share Mexico [Online].
› Ferrazzano G, Sangianantoni S, Mitrano R, Ingenito A, Alcidi B, Cantile Available: http://gs.statcounter.com/search-engine-market-share/all/
T. Assessing changes in oral health-related quality of life and body mexico [Accessed February 2019].
growth in 3-5 years old children following dental treatment under general › Statcounter. 2019 -b. Search Engine Market Share United States Of
anaesthesia due to severe dental caries. Eur J Paediatr Dent 2019; 20, America [Online]. Available: http://gs.statcounter.com/search-engine-
214-218. market-share/all/united-states-of-america [Accessed February 2019].
› Flesch R. A new readability yardstick. J Appl Psychol 1948; 32, 221. › Tan SS-L, Goonawardene N. Internet health information seeking and the
› GBD Oral Disorders Collaborators, Bernabe E, Marcenes W, Hernandez C, patient-physician relationship: a systematic review. J Med Internet Res
Bailey J, Abreu L, Alipour V, Amini S, Arabloo J, Arefi Z. Global, regional, 2017; 19, e9.
and national levels and trends in burden of oral conditions from 1990 to › Wineburg S, McGrew S. Lateral reading and the nature of expertise:
2017: a systematic analysis for the global burden of disease 2017 study. J Reading less and learning more when evaluating digital information.
Dent Res 2020; 99, 362-373. Teach Coll Rec 2019; 121, 1-40.
› Glick M. Believing is seeing: Confirmation bias. J Am Dent Assoc 2017; › Winker MA, Flanagin A, Chi-Lum B, White J, Andrews K, Kennett
148, 131-132. RL, DeAngelis CD, Musacchio RA. Guidelines for medical and health
› Greyson D. Information triangulation: A complex and agentic everyday information sites on the internet: principles governing AMA web sites.
information practice. J Assoc Inf Sci Technol. 2018; 69, 869-878. JAMA 2000; 283, 1600-1606.
› Ipsos 2018. Global views on healthcare: Percentage of adults worldwide › Yardi S, Caldwell PH, Barnes EH, Scott KM. Determining parents’ patterns
who used select sources for health information as of 2018; 84 of behaviour when searching for online information on their child’s health.
[Online]. Available: https://www.ipsos.com/sites/default/files/ct/news/ J Paediatr Child Health 2018; 54, 1246-1254.

20 European Journal of Paediatric Dentistry vol. 23/1-2022