Validation of a Persian Short-Form Version of a Standardised Questionnaire Assessing Oral Cancer Knowledge , Practice and Attitudes Among Dentists

OBJECTIVES
Oral cancer is a global health problem; however, many dentists lack the necessary skills, knowledge and capacity to diagnose oral cancers early. This study aimed to examine the validity and reliability of a Persian short-form version of a standardised questionnaire to assess dentists' knowledge, practice and attitudes towards oral cancer.


METHODS
This cross-sectional analytical study was carried out in May 2015 in Tehran, Iran. An original 39-item English-language questionnaire developed by Yellowitz et al. was translated into Persian using forward and backward translation methods. A total of 15 dental professionals were asked to assess the questionnaire for content validity. Based on their feedback, a 20-item short-form version was prepared, including six demographic, six knowledge, four attitude and four practice items. The translated short-form questionnaire was subsequently distributed to 973 general dental practitioners attending a dental conference in Tehran. Internal consistency and reliability were assessed with Cronbach's alpha coefficient and item-total correlation calculations.


RESULTS
A total of 13 professionals and 313 general dentists participated in the study (response rates: 86.7% and 32.2%, respectively). After the elimination of six items (two knowledge, two attitude and two practice items), the validity and reliability of the questionnaire was confirmed.


CONCLUSION
The final Persian 14-item version of the questionnaire had acceptable validity and internal consistency. These results indicate that researchers can use this translated short-form version to evaluate oral cancer knowledge, attitudes and practices among Persian-speaking dentists; this will allow for a comparison of data between different populations.

O ral cavity cancers are considered a major health risk, especially for individuals in developing countries; approximately 400,000 new cases of oral and pharyngeal cancer are identified annually worldwide. 1In addition, almost 13,000 deaths occur globally per year due to squamous carcinomas of the head and neck and there has been a relative increase in the incidence of oral cancer in recent decades. 2 Unfortunately, many oral cancer lesions are not diagnosed until they have reached advanced stages; for example, squamous cell carcinomas in the oropharynx may be confused with pharyngitis and viral tonsillitis, consequently resulting in delays in diagnosis and treatment. 3Since the majority of individuals undergo dental check-ups at least once or twice a year, dentists can carry out thorough and accurate examinations of regions at high risk for oral cancer, potentially allowing for an early diagnosis. 4][6] Previous studies have emphasised the importance of promoting knowledge of oral cancer among dentists so as to ensure rapid patient screening, successful treatments and decreased mortality rates. 5,6Therefore, the design and implementation of a standard validated questionnaire in this field is important; it would reduce bias in relation to the selection of questions, thereby allowing a more accurate comparison of results from different studies.Yellowitz et al. introduced the first comprehensive questionnaire to determine dentists' knowledge, attitudes and practices in relation to oral cancer. 7owever, many dentists did not have the time or inclination to answer all of the questions in the original 39-item questionnaire; the incomplete questionnaires resulted in a decrease in the accuracy of responses, thereby compromising the validity of the evaluation. 7The present study was undertaken to validate a Persian short-form version of Yellowitz et al. 's original English-language questionnaire on dentists' oral cancer knowledge, attitudes and practices. 7

Methods
This cross-sectional analytical study took place in May 2015 in Kerman and Tehran, Iran.An original English-language questionnaire for dentists developed by Yellowitz et al. on oral cancer knowledge, attitudes and practices was translated into Persian. 7Five items (questions 22, 25, 31, 37 and 39) in the original questionnaire were deemed to be specific to the USA and were removed at the agreement of the translators and researchers.The resulting 34-item questionnaire was first translated based on standard principles; as such, two experts in translating dental technical terms separately carried out a literal translation of the original English-language questionnaire into Persian. 8f the literal translation resulted in distortion of the meaning of a question, a conceptual translation was performed.Subsequently, two other individuals translated the questionnaire back into English; these individuals were experienced in translating from Persian into English and had not previously read the English version of the questionnaire.The resulting translations were then compared with the original text to identify possible discrepancies.Discussions were held among all of the translators until an agreement on the final text was reached.Final corrections were subsequently made to the questionnaire items with discrepancies. 9e final Persian translation of the 34-item questionnaire was distributed to 10 oral medicine and five oral/maxillofacial pathology specialists from Kerman Dental School, Kerman.In order to prepare a short-form version of the questionnaire, the specialists were asked to evaluate the questionnaire and identify items that could be removed.As per the feedback obtained, only items deemed necessary by >80% of the specialists were included in the shortform version of the questionnaire. 9This resulted in the final inclusion of six demographic and 14 oral cancer items (six knowledge, four attitude and four practice items).One of the questions eliminated from the final version was related to precancerous lesions.
Subsequently, a total of 973 general dental practitioners participating in the 55 th Annual Congress of the Iranian Dental Association in Tehran were recruited via a convenience sampling method to complete the Persian short-form version of the questionnaire.The necessary sample size was calculated to be 300 (α = 95%; d = 0.05; P = 38%). 10ompleted questionnaires in which over 30% of the questions were unanswered were not included in the analysis.
Data from the questionnaire responses were analysed using the Statistical Package for the Social Sciences (SPSS), Version 21 (IBM Corp., Chicago, Illinois, USA).Results were analysed descriptively using a Student's t-test and Chi-squared test.The internal consistency and reliability of the questionnaire items were determined using Cronbach's alpha coefficients and item-total correlations. 8,9Acceptable cut-off values for Cronbach's alpha coefficients and item-total correlations were set at ≥0.70 and ≥0.30, respectively.If an item resulted in an unacceptable coefficient value, the item was eliminated and the coefficient was recalculated; if the new coefficient was subsequently acceptable, the item was removed entirely from the final version of the questionnaire.For attitude items, items were scored on a five-point Likert scale, with scores of 1 indicating strongly disagree and 5 indicating strongly agree.
The protocol for this study was approved by the Ethics Committee of the Kerman University of Medical Sciences, Kerman, Iran (#K.95.62).Informed oral consent was obtained from all of the dentists before participating in this study.All personal and identifying information was kept confidential.

Results
The Persian short-form version of the questionnaire was completed by 323 general dentists; however, a total of 10 questionnaires were excluded from the study due to unclear responses and a lack of demographic data, resulting in a final sample size of 313 participants (response rate: 32.2%).Table 1 presents the demographic characteristics of the participants.No significant differences were found between demographic characteristics and oral cancer knowledge, attitudes or practices.With regards to the six knowledge items, the second item had the highest  rate of correct responses (66.7%).The elimination of any of these six items did not result in acceptable Cronbach's alpha values; however, items five and six were eliminated due to their low item-total correlation coefficients (0.05 and 0.12, respectively) [Table 2].For the four attitude items, calculation of the item-total correlation coefficients resulted in the elimination of items one and two [Table 3].In addition, the initial overall Cronbach's alpha value for all of the attitude items was 0.32; this increased to 0.62 (close to the cut-off value of 0.70) after elimination of these two items, indicating the necessity of their removal from the final version of the questionnaire.
In the practice section, elimination of none of the individual items resulted in acceptable Cronbach's alpha coefficients [Table 4].The overall Cronbach's alpha coefficient for all of the practice items was 0.11; however, following the removal of items one and two, this was recalculated to be 0.64 (close to the cut-off value of 0.70).In addition, these two items exhibited unacceptable item-total correlation coefficients (0.21 and 0.04, respectively).Therefore, these two practice items were subsequently eliminated from the final version of the questionnaire.An English version of the final short-form questionnaire is shown in Table 5.

Discussion
The current study sought to validate a Persian short-form version of an existing English-language questionnaire assessing dentists' knowledge, practice and attitudes towards oral cancer. 7The original questionnaire was translated from English into Persian and then back into English; however, this process can introduce errors.It is vital that the original intention of the author of a questionnaire item not be changed during translation, as the purpose of the translation process is not to rephrase or improve an existing instrument. 11Fortunately, the majority of the suggestions made by the specialist evaluators in the current study referred to the style of the translation  and no translation errors were reported.Overall, the Persian short-form version of the questionnaire exhibited suitable validity and reliability; this is consistent with the findings of other studies. 9,12The availability of high-quality translations is of vital importance in order to guarantee the successful implementation of existing questionnaires in other non-English-speaking populations and to assure the international comparability of the results.
][15] Consequently, this results in divergent scores, making it impossible to compare total scores between studies; for example, Razavi et al. reported an acceptable knowledge score to be ≥53.8%(seven or more correct responses out of 13 questions) among Iranian dentists, while Honarmand et al. reported poor knowledge scores to be ≤33.3%(less than four correct responses out of 12 questions) among senior dental students in Iran. 13,15Kujan et al. reported that 81% of Saudi undergraduate dental students correctly answered all knowledge questions, although the researchers did not define what constituted acceptable knowledge scores. 14n the current study, 26.7% of the dentists correctly answered the knowledge item related to the most common location of oral cancers; this is low in comparison to previous results from the USA and Iran (39% and 80.9%, respectively). 16,17[20][21] Various factors may affect differences in knowledge among dentists, such as potential differences in undergraduate dental curricula worldwide.A total of 33.3% of the dentists in the current study were aware of the predominant age range in which oral cancer presents.In comparison, Vijay Kumar et al. reported that 59% of Indian dentists correctly responded to this question. 22López-Jornet et al. reported that short-term oral cancer knowledge scores among the general population increased following an educational intervention. 23ith regards to knowledge of the metastatic lymph node characteristics of oral cancer, 48.6% of respondents in the current study provided correct responses.Decuseara et al. reported that 54% of Irish dentists in their study had been adequately trained to palpate lymph nodes to detect oral cancer. 24It is often difficult to determine which aspects of oral cancer are most important and should be prioritised in questionnaires assessing oral cancer knowledge.Common symptoms of oral cancer and awareness of diagnostic tools, such as cytology smears, are rarely emphasised. 14,16,20A major difference between the final short-form version of the questionnaire used in the present study and those of similar studies was the lack of inclusion of questions assessing awareness of the main risk factors for oral cancer (i.e.smoking and alcohol use). 19,22,25,26One reason for removing this  item from the final version of the short-form questionnaire was because it was deemed relatively easy in comparison to other knowledge-related questions.However, Horowitz et al. reported low overall knowledge of risk factors and signs or symptoms of oral cancer among adults in the general population. 27he inclusion of such questions should be re-evaluated in future studies.
With regards to attitudes towards oral cancer, 29.2% of Iranian respondents in the current study believed that they were well-trained in the performance of oral cancer examinations; this finding is much lower than that reported by Colella et al. (64.8%). 20Carter et al. found that significantly more general dental practitioners than general medical practitioners routinely inspected the oral mucosa during patient examinations (95.49% versus 20.17%; P <0.001). 28In the present study, 82.1% of the Iranian dentists expressed an interest in attending continuing education programmes regarding oral cancer; this is higher than that reported by Shaila et   19,31,32 Overall, few participants in the current study believed that their oral cancer knowledge was up-todate (26.7%).In a study by Decuseara et al., 53% of respondents provided positive responses to a similar question. 24Carter et al. also reported that significantly fewer general medical practitioners felt that they had sufficient knowledge regarding the prevention and detection of oral cancers in comparison to general dental practitioners (25.2% versus 54.1%; P <0.001); however, both general medical and dental practitioners requested further oral cancer training (71.4% and 80.4%, respectively). 28The majority of the dentists in the current study believed that annual oral cancer examinations were necessary for patients over 40 years old (82.2%).
Responses to questionnaire items in the practice section indicated that 94.2% of respondents in the present study referred patients with suspicious oral lesions to specialists; this was consistent with the findings of Vijay Kumar et al. and Dumitrescu et al. (98% and 92.9%, respectively). 18,22In addition, the majority of Iranian dentists in the current study questioned patients about their history of other cancers (75.9%), family history of cancer (75.2%) and smoking habits (87.2%).In a study by Rocha-Buelvas et al., 22.8% of South Colombian dentists asked their patients about a family history of cancer while Tanriover et al. found that 70.1% of Turkish primary care physicians asked patients about their tobacco use habits. 21,33Only 27.9% of Iranian dentists in the present study believed that they possessed the necessary training to instruct their patients on how to quit smoking; in contrast, Kujan et al. observed this rate to be 63% among Saudi dental undergraduates. 14everal studies have posed this question in different formats. 22,24,29,34It should be noted that 57% of the undergraduate dental students in Kujan et al. 's study believed that it was not their responsibility to encourage patients to give up smoking. 14In the final short-form version of the questionnaire designed in the present study, the questionnaire item on the number of biopsies taken was removed; moreover, no items on the number of oral cancer examinations performed were included.Previously, Shaila et al. reported that 29.5% of Indian dentists took biopsy samples of questionable lesions. 29Several previous studies have assessed the number of oral cancer examinations performed among dentists. 16,22,24,35uture research should seek to determine the necessity of including such questions in this type of questionnaire.
Determining the awareness, attitudes and practices related to oral cancer among dentists is critical; however, homogeneous study tools are necessary to ensure that the evidence and conclusions drawn from specific research is more widely applicable and comparable between population groups.In addition, short-form questionnaires may result in larger sample sizes, as participants are more likely to complete the entire questionnaire.However, the current study was subject to certain limitations.It is likely that those who did not answer all of the questions had poorer knowledge in comparison to those who did.As such, the knowledge levels of the dentists reported in the current study may have been inaccurate, hence weakening the validity of the results.Further studies are recommended to improve the potential weaknesses of this study.

Conclusion
The findings from this study indicate that the final Persian 14-item version of a standard questionnaire had acceptable validity and internal consistency.Researchers can use this short-form questionnaire to evaluate Persian-speaking dentists' knowledge of, attitudes towards and practices associated with oral cancer.This would subsequently allow for greater comparison of findings between different populations.c o n f l i c t o f i n t e r e s t The authors declare no conflicts of interest.

Knowledge 1 .Practice 7 .
The most common site of oral cancer besides the lips is the tongue and oral cavity floor 2. The most common form of oral cancer is squamous cell carcinoma 3. Oral cancer typically presents at 30-49 years of age 4. In oral cancer, metastatic lymph nodes are fixed and nontender, with a stony-hard consistency Attitude 5. a) My knowledge of oral cancer is up-to-date b) It is necessary to perform annual oral cancer screening examinations for patients >40 years old c) It is easy to refer patients with suspicious lesions to oral cancer specialists d) The early detection of oral cancer increases five-year survival rates 6. a) I am well-trained in providing education on smoking cessation b) I am well-trained in performing oral cancer examinations c) The majority of dentists can competently perform oral cancer examinations d) The majority of physicians can competently perform oral cancer examinations Referral department for patients with suspicious lesions: a) Oral medicine b) Other 8. Factors evaluated during history-taking: a) Smoking b) Alcohol consumption c) Smoking type and frequency d) History of other cancers e) Family history of cancer *An English-language questionnaire on dentists' oral cancer knowledge, attitudes and practices developed by Yellowitz et al. 7 † Not including six demographic items included in the final questionnaire.

Table 1 :
Demographic characteristics of general dental practitioners in Tehran, Iran (N = 313)

Table 2 :
Internal consistency and reliability of oral cancer knowledge items in a Persian short-form version of a standardised questionnaire* completed by general dental practitioners in Tehran, Iran (N = 313) 7-T corr = item-total correlation.*AnEnglish-languagequestionnaire on dentists' oral cancer knowledge, attitudes and practices developed by Yellowitz et al.7† Using correct responses only.‡ This value represents the Cronbach's alpha coefficient after elimination of the item from the questionnaire.

Table 3 :
Reliability of oral cancer attitude items in a Persian short-form version of a standardised questionnaire* completed by general dental practitioners in Tehran, Iran (N = 313) 7An English-language questionnaire on dentists' oral cancer knowledge, attitudes and practices developed by Yellowitz et al.7† Including positive attitudes only.

Table 4 :
Internal consistency and reliability of oral cancer practice items in a Persian short-form version of a standardised questionnaire* completed by general dental practitioners in Tehran, Iran (N = 313) 7An English-language questionnaire on dentists' oral cancer knowledge, attitudes and practices developed by Yellowitz et al.7†This value represents the Cronbach's alpha coefficient after elimination of the item from the questionnaire.
In contrast, Dumitrescu et al. reported that 54.3% of Romanian dental students were aware of this fact. 18Overall, 66.7% of respondents in the present study correctly identified the most common form of oral cancer; this is lower than the percentage of correct responses reported by Alaizari et al. (82.8%) and higher than those reported by Colella et al. (50.5%) and Rocha-Buelvas et al. (52.69%).

Table 5 :
Questionnaire items included in the final Persian short-form 14-item version of a standardised questionnaire* assessing oral cancer knowledge, practice and attitudes among dentists al. (72%) and Gajendra et al. (75%). 29,30Dentists have also indicated a need for more training to detect oral cancer, for example in studies by Ariyawardana et al. (70%), Jaber et al. (37.6%) and Alaizari et al. (86%).