A validated instrument to assess dentists’ knowledge about diabetes: the dental -diabetes questionnaire

Aim: To elaborate and validate an instrument for Brazilian Portuguese speakers, to assess dentists’ knowledge about care of patients with diabetes mellitus (Dental-Diabetes). Methods: Methodological study comprising four stages: a) Elaboration of instrument; b) Content validation (computing Content Validity Index CVI) based on Expert Committee assessment; c) Pre-test with 30 dentists, followed by assessment of suggestions by Expert Committee; d) Psychometric validation through instrument application in a sample of 127 dentists by means of the web tool e-Surv. Cronbach’s alpha and intraclass correlation coefficients were used to evaluate, respectively, internal consistency and reproducibility. Results: The final version of the instrument consists of 22 questions (7 on sociodemographic data and 15 querying dentists’ knowledge) and those submitted for validation attained a CVI of 0.95 [95% CI 0.916-0,981], showing satisfactory internal consistency, with 0.794 Cronbach’s alpha [95% CI 0.741-0.842] and an intraclass correlation coefficient of 0.799 [95% CI: 0.746-0.846] between the test and retest scores. Conclusions: Dental-Diabetes is a comprehensive instrument, culturally adequate and validated to assess dentists’ knowledge about care of patients with diabetes.


Introduction
Diabetes Mellitus (DM), a chronic disease, is a public health problem that affects a large number of individuals from all social backgrounds. In 2019, the International Diabetes Federation (International Diabetes Federation, IDF) estimated that diabetes affected 9.3% of the world population, with an estimated prevalence of more than 10.9% people in 2045 1 . If uncontrolled, systemic complications of DM can include heart attack, kidney disease, limb loss, blindness, and peripheral nerve damage 2 .
Due to the complexity of diabetes, how well the disease is controlled is an important issue in dental treatment planning. Patients with DM manifest a high prevalence of oral problems such as periodontal disease, tooth loss, xerostomia, caries, burning mouth disorder, taste and salivary gland dysfunction, delayed wound healing, lichen planus, geographic tongue, and candidiasis 3 .
Being a common disease in dental practice, dentists are expected to be part of a multidisciplinary team, collaborating especially with endocrinologists. They are expected to base their care on strategies to provide effective management of DM and its oral consequences, identifying oral disease impacting glycemic control, which, in turn, can impact oral health 4 . They are also expected to be aware of the pathophysiology of DM, its oral manifestations, signs and symptoms, how to react in case of an emergency, risks involved, systemic repercussions of the use of medications and anesthetics, which can all add up to provide better care for patients [5][6][7] .
Assessing dentists' conduct regarding DM is crucial to understand their knowledge, contributing to establishing targets for their training in public and private services and teaching institutions and defining guidelines for educational content and contributing to better care for patients with DM. Given the importance of dentists' role and the lack of tools to assess their knowledge about DM, an instrument was felt to be needed to gather information about dentists' knowledge and promote their education on related topics. A specific instrument for this purpose is not currently available. The aim of the study was to develop and validate an instrument for the assessment of dentists' knowledge about diabetes (Dental-Diabetes).

Materials and Methods
This is a methodological and exploratory study carried out from January 2017 to August 2018 in the city of Belo Horizonte, in the State of Minas Gerais, Brazil. The project was approved by the Ethics and Research Committee Involving Human Beings (CAAE number 65656117.6.1001.5138) at Santa Casa of Belo Horizonte Hospital. Agreement to participate in the study was obtained by using a Free Informed Consent Form signed by participants when accessing an electronic questionnaire by means of the webtool e-Surv. An Expert Committee made up of five dentists, a nurse, an endocrinologist, a linguist and a statistician took part in elaboration of the instrument and assessed all stages until its final version ( Figure 1).

STAGE 1 -Instrument's development
Prior to elaboration of the instrument, three authors (M.A.G.M.G., A.A.O.P., J.S.R.) conducted a literature review in PubMed (U.S. National Library of Medicine), LILACS (Latin American and Caribbean Literature in Health Sciences databases), and SciELO (Scientific Electronic Library Online) databases to obtain state-of-the-art information about diabetes, national and international recommendations on dental treatment of people with diabetes and questionnaires used to assess professionals' knowledge about a particular disease. The descriptors used for the queries were 'Diabetes mellitus', 'Dentists', 'Knowledge', and 'validation studies'. National and international publications yielded by the database queries were screened 2,5-13 . Based on the gathered insights we decided to elaborate an instrument in the form of a questionnaire.
In establishing a general conceptual structure dentistry-endocrinology interface, our instrument was developed in two parts: the first section focusing on dentists' socio-  demographic profile (7 questions); and the second section aimed at assessing dentists' knowledge about key aspects of DM and related care expected to be performed as part of their work (16 questions) (Version 1-V1).

STAGE 2 -Content validation
For content validation, a web address to access a web assessment form was sent by e-mail to 22 dentists, six endocrinologists and five linguists (Judges' Committee), who evaluated each item of the instrument's first version (V1). Criteria for participation in the Committee were either to be a professional dentist with or without clinical practice implicating diabetes (dentist profile); or have taken part in questionnaire elaboration or translation in the healthcare area (linguist profile); or to have clinical practice in diabetes (endocrinologist profile). Judges were selected based on their curriculum vitae.
The Judges' Committee assessed clarity and relevance of each item in V1 and rated them with the following options: one star standing for need for full reformulation; two stars, partial reformulation (substantial revision needed); three stars, need for partial reformulation, with minor editing to enhance text style; and four stars in case of no need for reformulation. A comment box was also provided for the experts' remarks and suggestions.
Once the evaluation was completed, the Content Validity Index (CVI -the level of agreement of experts on adequacy of the items) was computed: number of scores 3 and 4 divided by total number of scores by all Committee members. CVI indicates the degree to which a scale has an adequate sample of items to represent a construct of interest -that is, whether a domain of content for the construct is adequately represented by the items. Results higher than or equal to 0.78 are considered acceptable 14 . Upon computing CVI (higher than 0.78) and implementing the Committee's suggestions to improve, V2 was obtained. V2 is substantially similar to V1, except for minor editing and spelling correction.

STAGE 3 -Pre-test
The author (M.A.G.M.G.) carried out the pre-test through face-to-face interviews with 20 dentists 15,16 in 10 meetings following participants' schedules -2 group meetings with 5 dentists, 2 pair meetings and 6 individual meetings. First, the whole instrument was read by each participant individually; secondly, items were discussed to ensure whether they were clear, accurate, relevant and adequately arranged. Participant's feedback was then assessed by the Expert Committee, who considered all relevant comments and redrafted those items that obtained less than 80% agreement 14 . Version 3 (V3) was thus obtained and tested on a newly selected group of 10 dentists (in 8 individual meetings and 2 in pair meetings).

STAGE 4 -Validation
A web address to access V3 in digital format on the e-Surv platform was sent via e-mail to 127 dentists selected by convenience from both public and private services and universities. The sample size was adequate considering a level of significance equal to 5%, test power equal to 80%, standard deviations equal to the test and retest scores and a correlation coefficient equal to 0.30 (minimum value detected in the consistency assessment). A minimum sample size requirement was 85 professionals. Retest was performed with those 127 dentists with a minimum interval of 7 days and a maximum of 21 days between the tests (average 16 days) 17 .

Statistical analysis
Absolute and relative frequencies were used to describe the sample characteristics and the proportion of correct answers to the instrument items. Internal consistency and reproducibility were verified to analyze the reliability of the construct. Cronbach's alpha (CA) was used to assess the internal consistency of the instrument. Internal consistency is an assessment of whether items intended to measure the same construct produce similar scores. A high degree of internal consistency indicates that items meant to assess the same construct yield similar scores. There are a variety of internal consistency measures. Usually, they involve determining how highly these items are correlated and how well they predict each other. Cronbach's alpha is a commonly used measure.
The instrument's reproducibility was evaluated through test-retest (temporal stability), computing the intraclass correlation coefficient (ICC). The Kappa index was added for reproducibility and refers to the percentage of concordant responses in the test and retest, defined as the ratio between the number of individuals who selected the same answer (regardless of being correct or incorrect) at both test and retest and the total number of individuals 16,18 . Floor and ceiling effects were measured by the number of respondents receiving the minimum and maximum scores, respectively. The significance level adopted for the statistical tests was 5%. For data analysis, SPSS version 20.0 was used.

Instrument development and Content validation
The development spanned 6 months. V1 of the instrument consisted of 23 questions. After reviewing V1 following the experts' suggestions, a second version (V2) was obtained, with 23 questions. The instrument achieved a good score by the committee regarding clarity and relevance, with a total CVI of 0.95 [95% CI 0.916-0,981].

Pre-test
In face-to-face tests, which lasted 3 months, 30 dentists participated. 67% were female; 47% had a Diploma course and were working in different areas such as surgery, dentistry, endodontics, periodontics and others; 47% had more than 20 years' experience; 100% reported having provided dental care to patients with diabetes; 67% declared not having had any training to treat patients with diabetes; however, 63% reported feeling empowered to provide care for Diabetes patients. (Table 1).
As an outcome of the first meeting, with suggestions by 20 dentists, 2 questions were merged in order to adapt terms and increase understanding, a total of 22 items remaining in the questionnaire, yielding a third version (V3). V3 was tested with ten other dentists, no need for further redrafting having been requested. V3 was hence considered adequate to be submitted to psychometric validation. The 15 questions on knowledge about diabetes were then submitted for validation (Table 2).

Validation
This stage lasted about 28 days and 127 dentists answered the final version of the instrument (test and retest) ( Table 1). 70% were female, 53.5 % had a Diploma course, 40.9% had expertise in general clinical practice., and 52.8% had obtained their first degree 20 years ago. 94.5% had already provided dental care to people with diabetes.
The total CA alpha value was 0.794 (95% confidence interval 0.741-0.842). Floor effects (percent with minimum score) were 0%, and ceiling effects (percent with maximum score) 3,2%. An ICC value of 0.799 (95% CI: 0.746-0.846) was obtained. The Kappa coefficient, which assesses the degree of agreement, varied between 0.5-1.0 (mean: 0.80). When the alpha absence index was calculated, there was a slight impact on reducing AC and no questions needed to be removed (Table 3).

Discussion
The treatment of patients with diabetes requires knowledgeable professionals, dentists being fundamental member in a multidisciplinary team; therefore, dentists are expected to be updated regarding diabetes and its implications for daily care, with a greater knowledge about the onset, duration and control of the disease, resulting in a more effective and satisfactory approach 19 . A good interaction between the dentist and the multidisciplinary team is essential for a safer dental treatment, with lower chances of complications for the patient 20 .
Given the importance of the dental approach for the patient with DM, elaborating and validating an instrument to evaluate the dentists' knowledge about DM was considered important to identify possible flaws in the knowledge of diabetes of these professionals that could impact treatment decisions and the objectives of the patient. In Brazil, dentists are not yet part of teams in diabetes centers as is the case in other countries, despite the clear need for these professionals to share the knowledge and duties of a multidisciplinary team. The collaborative work by the Expert Committee pooling expertise in diabetes, dentistry and language issues made it possible to elaborate a comprehensive instrument, solving problems encountered during the process of drafting and adapting concepts and terms to the language used by the target subjects 13,16,18,[21][22][23] . Interaction between healthcare professionals and applied linguists is a fundamental piece in the elaboration and cultural adaptation of new instruments.
Assessment by the Expert committee through the web tool e-Surv is a reliable and efficient methodology 24 , allowing for remote application and quick data extraction, avoiding potential errors in transcriptions and gathering of results 25 . The instrument successfully passed the Committee's examination in terms of clarity and relevance, with an excellent CVI (0.95). It should be noted that the maximum value for CVI is equal to 1, the results achieved being close to the maximum score 18 , well above the CVI cutoff point of 0.80 for new instruments 18,26 .
In the pre-test, the face-to-face meetings with a sample of dentists proved successful, favoring adjustments in the instrument, and ensuring the prospective understanding of the items by the target audience 27,28 . In carefully developed instruments, two or three face-to-face tests can be satisfactory, which was the case in our study, with two rounds being required 21 .
For validation (test-retest), a CA index of 0.794 was obtained, which indicates good internal consistency [29][30][31][32] . The time span for retest met the recommendations in the literature: a 7 to 21 day interval (a mean of 16 days). There are controversies regarding interval between test and retest, a desirable interval being not too short for participants to recall their answers in the test and too long for the study to be impacted [33][34][35] .
There are recommendations of an interval of one to two weeks between test and retest; however, no fixed amount of time is prescribed, the main concern being the need to account for whatever interval span chosen 36 . The time interval in our study adhered to the above recommendations, variation being due to participants' agendas.
Our CA, ICC (0,799) and Kappa index (mean 0.80) indicate that our instrument showed adequate stability, reproducibility and confidence 18 . ICC being satisfactory, we computed Kappa to corroborate it. Items with a low Kappa (5a, 5c e 5d, 12b e 12c) revealed topics that were less familiar to dentists.
CA absence index was carried out. Removing items (1, 5e, 6, 11 e 12a) yielded alpha scores higher than those for the whole set of items. Therefore, no questions were excluded, due to the small difference that would result in the final CA 37 and the possibility of leaving out important information 13 . In addition, the value was above 0.799 for all items and thus considered satisfactory.
When we analyzed the performance of dentists in the test, the questions that had the lowest percentage of correct answers were question 2 (time period considered in the glycated hemoglobin test to assess mean blood glucose levels); 5 (hypoglycemia and signs of mood change and/or irritability); and 14c (use of sedatives). This percentage of incorrect answers was somehow predictable, since those are the most common questions asked to endocrinologists by dentists before dental procedures in daily care. In this respect, it is worthy of note that both in the pre-test and the validation stages, the majority of participants had over twenty years' experience and Braz J Oral Sci. 2022;21:e225337 had provided dental care to patients with diabetes; nevertheless, most had had no training whatsoever in diabetes care. Still, despite the lack of training, most reported feeling confident to treat patients with diabetes. This finding reveals a major problem in diabetes education and clinical practice regarding a highly prevalent condition as is diabetes and showcases the need for questionnaires such as the one we have elaborated and validated in our study as an instrument to assess the level of knowledge about diabetes by dentists and propose educational initiatives to contribute to better dental treatment for people with diabetes.
No studies were found describing the development and validation of instruments to assess dentists' knowledge about diabetes, which did not allow for our results to be compared. Our instrument comprises items implicating knowledge compatible with themes indicated as priorities for dentists' care of patients with diabetes: diagnostic criteria, symptoms, urgency and emergency, dental risk, conduct in care, clinical signs in the oral cavity and use of anesthetics and medications.
In conclusion, our study yielded an instrument that proved useful, reliable and stable for use by dentists. The instrument is useful to evaluate dentists' knowledge and promote professionals' training, with potential impact to enhance treatment for people with diabetes.