Background

Oral health has been growing and strengthening itself as an integral and inseparable part of public health policies, based on the guidelines proposed by the Brazilian Ministry of Health, by means of a model of care that seeks to expand access to services through the Oral Health Teams (OHT) included in the Family Health Strategy (FHS) [1, 2]. Brazil is a large country with great territorial and populational extension (approximately 210 million inhabitants), and it is the only country in the world with more than 200 million inhabitants to include the right to health as constitutional and universal. However, it is known that equalizing access and guaranteeing the integrality of actions is only possible through the organization of care based on population management [3, 4].

Despite the expressive increase in the implementation of OHT within the FHS over the years, i.e., superior than 470% within a 14-year period, and the growth in population coverage, 57% of the population still has no access to dental care. Therefore, an instrument that enables identification of the primary sources of vulnerability amidst the reality of millions of people can change this scenario, strengthen work management, and propose an equanimous offer in the access to oral health [3].

Nevertheless, equity in access remains a challenge to be resolved, in the context of considerable limited resources and a repressed demand, with services seeking strategies to prioritize individuals with the greatest needs in oral health [5, 6]. For the identification of oral health needs, it is essential to discuss the concepts of risk and vulnerability, two distinct concepts that are commonly addressed together [7,8,9].

The definition of risk is related to the probability of hazard, while vulnerability is defined as a characteristic of who or what is vulnerable or is fragile [10]. Vulnerability develops from the grouping of multiple risk factors that mutually reinforce each other, instigated by the lack of material, social, and environmental resources, which are indispensible for the individual’s well-being [11]. Some factors, such as age, ethnicity, working status, family income, schooling level, conditions of access to healthcare services, living conditions, and others, determine the individual’s greater or lesser susceptibility to harm and, consequently, higher or lower vulnerability [12]. The lowest vulnerability is represented by resilience and resistance of the individuals and communities when exposed to dangerous processes or events, by the so-called protection factors, such as a high educational level, economic stability, and increase in physical activity, among others [10, 12].

An association between vulnerability and oral health morbidity has been found in the poorest and most vulnerable regions of Brazil. Thus, the ability to measure vulnerability is recognized as an important element for the effective reduction of risk, and the use of instruments such as assessment scales could lead to a more equitable access of priority services [7, 10, 11], helping with the organization of the healthcare system.

Therefore, the use and impact of these types of questionnaires and scales have been evaluated before by Coelho and Savassi Scale [13,14,15,16], with the aim of prioritizing home visits, or by other authors [5], as a programming tool for the family's oral health teams, based on the social information (type of housing and maternal education) found in the Primary Care Information System.

This type of assessment instrument should measure or evaluate an object, so it is important that they have accumulated robust evidence of validity, and they should present a simple, clear language and cultural equivalence [17]. No scientifically validated instrument capable of measuring dental vulnerability was found in the literature [18]. Hence, this work aimed at developing and searching for evidence of content validity and internal structure of the Dental Vulnerability Scale (EVO) for PHC considering the Brazilian reality.

Methods

This is a psychometric study focused on the evaluation of evidence of content validity and internal structure, following current recommendation of the American Educational Research Association (AERA), American Psychological Association (APA), and the National Council on Measurement in Education (NCME) [19]. This study was approved by the Research Ethics Committee of the Municipal Health Department of São Paulo (Protocol 97371218.4.0000.0086).

The study was conducted in two stages, the first to evaluate content validity, and the second the internal structure of the DVS. Since there is no clear definition for vulnerability, initially a qualitative exploratory analysis was made to detect factors that could be associated with vulnerability, which in turn subsidized the development of the items of the instrument. According to DeVellis [20], a loss of 50% of items is expected during the validation process, therefore it was opted for an extensive stage to formulate the items.

Stage 1: Evidence of content validity

This stage utilized a convenience sample composed of 188 specialists (dental surgeons, oral health technicians, oral health assistants, undergraduate dental students, physicians, nurses, licensed practical nurses, and social workers) from all regions of Brazil, of both sexes, aged over 18 years, who digitally signed the Informed Consent Form, and filled out the research questionnaire received, which had as key question, “if you have to measure dental vulnerability, which risk factors would you consider?”. At this stage, content analysis was applied with the purpose of synthesizing the various meanings reported by the respondents. Bardin’s [21] methodological orientations were used.

The questionnaire was sent to specialists through a RedCap® link, reported via WhatsApp® application and by email, to groups that work directly or indirectly with oral health. The link sent could be forwarded to other specialists, the snowball method [22] was used to reach a larger number of participants in this stage of the study, with original recipients of the RedCap® being able to forward it.

To ensure reliability of the classification performed through the content analysis, four external and impartial judges were invited to review credibility of the categories established by the researchers. The items developed from the previous stage were submitted to a panel of 40 judges, who were asked to evaluate the relevance (item that checks dental vulnerability) and the clarity of the items (if the item was correctly written), besides allowing possible suggestions for improving the wording of the items. The Content Validity Ratio (CVR) developed by Lawshe [23] was applied to the analysis of the responses of the judges. Adoption of the CVR aimed at a more rigorous application of the content validity index, and allowed the adoption of a larger number of judges. The recommendation by Lynn [24] indicates that a judge panel size of five would be adequate, however more recent recommendations by Epstein et al. [25] specifies that the panel of judges should be composed of researchers, translators, professionals, methodologists, and lay persons, while Gilbert and Prion [26] emphasizesg that such panels should have specialists from different professional levels. Taking in to account these more recent recommendations, and the fact that Brazil is a country of continental dimensions, with a wide cultural diversity, it would have been risky to restrict opinions to only five experts, which could have biased the analysed.

Stage 2: Evidence of internal structure validity

The scale was answered by dental offices, in other Primary Care Units facilities, and in other home visits by field researchers trained at the administrative districts of Vila Andrade and Campo Limpo, located in the South Zone of São Paulo, with a estimated population of 387,408 individuals with 75% covered by the FHS. According to a survey conducted by the SEADE Foundation in 2019 [27], and data from the SIAB [28], 54.4% this population do not have private health insurance, and 58.9% depends exclusively on the Unified Healthcare System (SUS) [Brazilian National Public Health System].

To characterize the population studied, the “Brazil Classification Criterion” [29] and the individual and family registration forms of the E-SUS [30], were applied, in addition to the ICF.

Sample

The sample size of psychometric studies is usually estimated based on the number of items [31, 32] that demonstrate a ratio of 20:1 or more; that is, 20 respondents for each item of the instrument would be the ideal [33]. In this sample, a ratio of 29.9 persons/item was established, which preliminarily ensured that the sample size would be not responsible for possible inaccuracies in the analyses.

Statistical analysis

Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA)

Performance of the exploratory and confirmatory factor analyses requires the fulfillment of various stages, such as data inspection, factor analysis, factor retention technique, factor rotation technique, and factor load cut-off level [34]. These stages are systematically presented for validation of internal structure.

The dimensionality of the instrument was tested by applying the Robust Parallel Analysis (RPA) through Optimal Implementation of Parallel Analysis (PA) [35] with the Minimum Rank Factor Analysis (MRFA) that minimizes the common variation of the residues [36]. The robustness of the test was determined from the association of a bootstrap with an extrapolation of the sample to 5000.

The Parallel Analysis is considered one of the most robust and precise techniques used to test dimensionality [37,38,39]. The factors were extracted with robust Unweighted Least Squares (ULS) that reduce residues of matrices [31]. Non-orthogonal Promax rotation was applied [40].

The adequacy of the models for the exploratory and confirmatory factor analysis was evaluated taking into consideration the number of participants and the levels of the parameters that indicate an excellent explanation of the model. Thus, the minimum factor loads must be above 0.30; communalities with values above 0.40, and explained variance close to or greater than 60% [33, 41]. The maintenance or removal of items was done considering the explained variance of the model; factor load values and communalities below the minimum established.

For the CFA, the maximal likelihood method of extraction was adopted as a reference for model quality, using the cut-off techniques established by Sivo et al. [42]. The indices Goodness-of-Fit (GOFs): robust Mean and Variance-adjusted Chi-square/df (X2/df < 5) [43]; Non-Normed Fit Index (NNFI > 0.93); Comparative Fit Index (CFI > 0.94); Goodness-of-Fit Index (GFI > 0.95); Adjusted Goodness-of-Fit Index (AGFI > 0.93); Root Mean Square Error of Approximation (RMSEA < 0.07), and Root Mean Square of Residuals (RMSR < 0.08) were adopted [42].

Reliability

Three indicators were used: Cronbach’s Alpha [44], McDonald's Omega [45] and the Greatest Lower Bound—GBL [46]. The implementation of multiple reliability indicators aimed at reinforcing the model adjustment. In addition, various previous published articles have shown the inconsistency of reliability by means of Cronbach’s Alpha [47,48,49,50]. Alpha values greater than 0.90 should not be sought, since they can indicate duplication of content of the items and point towards redundancy more than to homogeneity [51]. Therefore, the authors chose analysis of three reliability techniques to increase precision of the study. Analyses were performed with the SPSS 23 [52] and Factor 10.10.03 [53, 54] software.

Results

The results are reported in three stages: development of the items, evidence of content and internal structure validity.

Stage 1: Development of the items

In the development of items for the EVO instrument, a primary questionnaire was sent to 188 oral health specialists. These experts, were mostly female (78.7%), dental surgeons (86.2%), having completed specialization degree (51.6%), and having professional experience at a PCU (79.8%), followed by work in a private office (73.9%). In order for the professional experience to be considered as adequate, this item could be marked more than once. All regions of Brazil contributed with answers, but the most expressive was the Southeast region (62.8%).

Analysis and classification of the content of the specialists’ responses for the qualitative exploratory stage resulted in the appearance of 172 factors associated with dental vulnerability, divided into 6 preliminary dimensions (healthcare services, social services, as well as services for mental health, oral health, infrastructure, and overall health). The factors were transformed into items of the instrument and prepared in an interrogative manner that enabled responses only in the dichotomous model (“0—No” and “1—Yes).

Stage 2

Stage 2.1 Evidence of content validity

After the development of the items, content validation was performed by 40 judges who gave their opinion on each item considering clarity, precision and relevance, and decided whether these items were suitable or not to compose the EVO. The selection of the 40 judges was made using the snowball method. The judges 60% were female, most resided in the Southeast region (92.5%), having their specialization as highest level of education concluded (52.5%). This panel was more heterogeneous, because in addition to the professionals linked to oral health (80%), other professional categories also participated, represented by physicians, nurses, speech therapists, pharmacists, and psychometrists.

The CVR establishes its critical value based on the number of judges. Therefore, for 40 judges the critical value was 0.25 [55,56,57]. The critical value of CVR was applied to evaluate the relevance (if the item measures the latent variable) and clarity of the items (semantics). The items that were at the critical value limit, that is, that presented with values of 0.25 for relevance, were included for the internal structure stage. The analysis resulted in the continuance of 41 of the 172 items initially developed. The result was a CVRmean of 0.33 for relevance of the item, and of 0.39 for the clarity of the item. Table 1 presents the values of CVR for relevance and clarity of the items that surpassed the critical value.

Table 1 CVR values for the preliminary version items
Full size table

Making an item clear is fundamental for a good understanding of the researched phenomenon, and it is necessary to structure concise writing, avoiding tautology to not lose relevance [58]. All items considered relevant (with CVR > or equal to 0.25) were included in the preliminary version, and items with less than 0.25 of clarity were changed in the text, as suggested by the judges. Some items that had suggestions from judges and had a CVR above critical value were also changed, so that they were clearer and more understandable to the target population.

Among the changes, there is “are you diabetic?” for example, considered relevant by the judges with a CVR = 0.40, presented with a clarity CVR of 0.25, requiring a reconfiguration of the writing so that it could be part of the items sent to the field. The item previously written as “are you diabetic” was changed to “do you have diabetes”, following a suggestion of the judges to modify the writing. In the same way, the item “do you have access to fluoridated water?” with a relevance CVR of 0.30 and a clarity index of 0.25, was revised and according to content analysis, had the writing changed to “do you have access to water with fluoride?”. Another example of an item considered as showing lower clarity (CVR = 0.10) and good level of relevance (CVR = 0.30) had the objective of knowing if the study participants were bedridden or not. The question initially written “Do you live limited at home?” was modified to “Do you live only inside the house?”.

Hence, the 13 items that needed rewriting for adequacy of the clarity item (CVR < 0.25) were revised and inserted into the final field version of the proposed scale.

Stage 2.2: Evidence of internal structure validity

The instrument was applied to 1227 participants in the South Zone of São Paulo, in the districts of Campo Limpo and Vila Andrade, at the PCU that are part of the partnership that the Sociedade Beneficente Israelita Brazileira Albert Einstein (SBIBAE) has with the Municipal Health Department of São Paulo (SMS/PMSP) since 2001 for implantation, implementation, and management of FHS units and teams. There are 30 OHT, 30 dental surgeons, 23 oral health technicians, and 30 oral health assistants.

The instrument was applied by the OHT professionals, duly and previously trained on the objective and methodology of the research. Additionally, a Whatsapp® group was created with all researchers involved to answer questions, ensure a higher standard, and minimize inter-examiner errors. Data were collected simultaneously at 11 PCU by 30 OHT, in the period from September 27, 2019 to November 28, 2019, during the work routine of the professionals (scheduled visits, dental emergencies, and home visits).

Stage 2.3: Characterization of the participants

As per Table 2, the instrument was answered by the largest number of participants possible within the determined period for collection at the primary care units. The respondents had a mean age of 33.30 years and a standard deviation of 17.30, and there was a prevalence of females (68.2%) and brown skin color (52.9%). The majority of these responded to the questionnaire during their routine dental visit (65.2%), while the other forms of capture from the respondents were distributed among urgent consultations (20%), home visits (3.6%), triages (4.9%), groups (1.8%), and others (4.5%).

Table 2 Profile of respondents
Full size table

As to the level of education, we noted that most of the participants had concluded high school (33.6%) or had not completed elementary education (32.9%); a small percentage had a salaried job (26.4%), while another portion was unemployed (23.6%). This scenario mostly pointed towards a family income of less than 2 minimum wages. As to living conditions, the majority of homes were composed of 4 rooms, and a median of 3 persons per household. In reference to oral health, the sum of teeth that were decayed, missing, and filled teeth (DMFT) had a median of 11.00 for adults, and when the same index for the adults was applied to children, which measures teeth that are decayed, extracted, and filled (DEFT), had a median of 3.50.

The initial analysis indicated that the measurements of adequacy of the sample established Kaiser–Meyer–Olkin (KMO) test of 0.65 and Bartlett test of sphericity of 8622.6 (df = 820; p < 0.01). The results indicate the data are adequate to be analyzed by factor analysis [33]. The correlations among the items were established between − 0.24 and 0.50, with a predominance of weak correlations, which can imply possible problems with adjustment of the model in the factor analysis.

For the initial test of dimensionality by the Parallel Analysis, the model proposed in content analysis had six dimensions. Dimensionality results indicated the existence of four dimensions for the set of items analyzed. The Parallel Analysis indicated relevant eigenvalues from 3.02 to 1.30, generating an explained variance of 36.56% for the model. Besides the dimensionality testing not confirming the six-factor criterion, only 26 out of 41items presented with factor loads higher than 0.30, and it was not possible to find an adequate alignment of these loads with the content of the initial factors. Only 13 items presented communalities superior to 0.30 (Table 3).

Table 3 Factor loads and communalities of the initial extraction with six dimensions
Full size table

Some GOFs of the CFA did not converge: chi-square, NNFI, CFI, and RMSEA; GFI = 0.96, AGFI = 0.95 and RMSR = 0.03 were established at appropriate levels. Nonetheless, several authors [59,60,61] pointed out that the presentation of the GOFs is not sufficient for the adjustment of the model. It should be clear that the factor loads, both by the confirmatory factor analysis (CFA) and the exploratory factor analysis (EFA), showed no adjustment of the model.

The inter-factor correlations remained between − 0.07 and 0.26, indicating a weak association among the dimensions of the instrument. The reliability values of the initial model, which were α of 0.58 and ω of 0.76, and GLB of 0.79, also did not show good levels. These results lose their relevance for not having model adjustment.

Faced with the non-adjustment of the model, the literature recommendation was adopted to remove the items that did not present adjustment of the model. For this, the values of factor load, communality, and cross-loading were adopted as criteria for maintenance or withdrawal of the item [33], in addition to the item content. The items were removed one by one and the analyses were redone until there was an adjustment of the internal structure, considering the statistical and content standpoints.

After removal of the items that did not reach the set of established indicators, the parallel analysis indicated a new model with four dimensions, with an explained variance of 72.11%. Measures of adequacy of the sample were also improved with the removal of items with KMO of 0.70 and Bartlett test of sphericity of 5295.1 (df = 105; p < 0.01). Table 4 displays the adjustment of factor loads and communalities of the new model.

Table 4 Factor loads and communalities of the final model with four dimensions
Full size table

Saturation of factor loads in Dimension 1 varied from 0.45 to 0.83, respectively, for items 5, 9, and 10. The contents of the saturated items in this dimension are coherent, which enabled the denomination of “overall health dimension”. In items 24, 25, 26, and 27, the saturation varied between 0.34 and 0.61, denominated from the “oral health” content. The “infrastructure” dimension was defined for the grouping of items 31, 32, 33, and 34 with saturation levels between 0.36 and 0.97; while items 36 to 39 belonging to the “healthcare service” factor had a saturation of 0.32 to 0.74. The adjusted model presented no cross-loading problems and Heywood cases. Therefore, all the items of the instrument substantially saturated in a single dimension and there was no violation of the limits (− 1 to + 1) in factor loads. The confidentiality values of the initial model were α of 0.64, ω of 0.99, and GLB of 0.82. Some items in the final version had communality values below the cut-off level, but they remained because the model proved adequate for all the other indicators.

The GOFs of the confirmatory factor analysis indicated values of × 2/df(51) of 3.23; in that, NNFI of 0.91; CFI of 0.94; GFI of 0.96; AGFI of 0.94; RMSEA of 0.04, and RMSR of 0.03. Besides presenting the GOFs that met the established criteria, the final model presented satisfactory values of the factor loads and predictive power, shown in Fig. 1.

Fig. 1
figure 1

Pathy diagram of the final model of the dental scale

Full size image

Final vulnerability scale

The final model presented factor loads for the confirmatory analysis that varied from 0.37 to 0.96, with R2 values that ranged from 0.14 to 0.91. Therefore, both the primary indicators (factor loads and predictive value of the item) and the GOFs were established at satisfactory values. In this way, the indicators, both of EFA and CFA, assure a good solution for the 15-item model with four dimensions presented in Table 5.

Table 5 Final dental vulnerability scale
Full size table

Discussion

The Dental Vulnerability Scale developed based on evidence of content and internal structure validity, showed four dimensions (overall heath, oral health, infrastructure, and healthcare services) and 15 distributed items.

Usually the development of a measuring instrument is based on theory/concept [62] and since the literature still has no instrument capable of exclusively measuring dental vulnerability [63], we adopted the concept defined as being a set of factors of the social, structural, general, mental and oral dimensions, in addition to health services and public management that influence the dynamics of the health-disease process in dentistry. The DVS proposed corroborates the concept described above, since the individual components are represented by the dimensions of overall health and oral health; the social or collective component by the dimension of the infrastructure; and lastly, the institutional component by the dimension of healthcare services.

The scales are developed when one seeks to measure a phenomenon that is not directly quantifiable, but a theory or concept points to its existence [20]. Certain indicators related to items of the scale allow obtaining a reasonably precise measurement of the phenomenon [33]. Therefore, the validation of a scale, which is a process in which we gather and evaluate the evidence to support the adequacy, significance, and usefulness of decisions and inferences, will validate the scale when it truly measures what it propose to measure, and for this, we seek the evidence of validity of the instrument items [33, 64].

Some instruments are widely used in Brazil, as the Oral Health Impact Profile (OHIP) [65], and the Coelho and Savassi Family Risk Scale was developed with the purpose of organizing home visits, based on the attribution of scores for sentinels considered relevant that were present in the SIAB. Despite presenting a sensitivity for defining families that require greater health care [16], this instrument has some limitations, such as difficulty in obtaining updated information, since studies have confirmed that updating at SIAB does not occur regularly [66], besides considering the risk for the family and not the individual, which can place a person within conditions not belonging to them. Other limitation is the scale does not have structure validity, and therefore, we cannot affirm if its items correlate, and if the scores are attributed in an adequate manner. Additionally, the sentinels encompass many families, classifying many of them as high risk, which hinder collection of concrete vulnerability data. Another issue is that there is no information about oral health in this scale; thus, we cannot use it to measure dental vulnerability [16, 66].

Studies addressing individual risk relative to oral health are scarce in the literature, making a comparative analysis difficult. This is the first study to prepare an instrument capable of measuring dental vulnerability, with characteristics that cover the entire national territory and with sufficient evidence of internal structure validity.

Evidence shows the instrument is multidimensional, comprising overall health, oral health, infrastructure, and healthcare services dimensions. As to the variables observed in the instrument after confirmatory factor analysis, we observed a greater relation between the “infrastructure” and “healthcare services” dimensions, and that the “overall health” dimension proved to be independent. This study showed that all proposed dimensions, present in the final fifteen items, may be appropriate to measure dental vulnerability in individuals from all Brazilian regions.

As to the items of overall health present in the DVS, it is necessary to ponder over the principle of integrality of individual care in the process of work, aiming at production of knowledge and new practices required in their care, for example, in chronic diseases, such as diabetes, hypertension, and obesity, which are very prevalent in the Brazilian population [67].

To consider systemic monitoring conditions as items of the scale is consistent with the literature. Obesity, for example, is a risk factor for innumerable systemic diseases, and there is a possible association between obesity and oral diseases [68, 69]. In the same way, diabetes, a morbidity that requires medical and dental follow-up, is related to dental vulnerability [70].

Additionally, hypertension, which is a chronic multifactorial disease that affects approximately 1.5 billion persons all over the world, appears or worsens in association with modification of oral conditions (gingivitis, periodontitis, and tooth loss) [71,72,73]. On the other hand, only clinical evaluations are not sufficient to measure health of the individual and identify vulnerable population groups, reinforcing the importance of using instruments that rate the perception of the individuals of their health and oral health [74].

The items related to oral health deal with the individual’s self-perception, which is determined by a need perceived as a result of their health condition and previous knowledge of the disease [75]. It can be considered a subjective indicator of the oral health condition that is strongly associated with the pattern of demand for dental services and the definition of priorities for planning, aiming to reduce inequities [76, 77]. When one observes the characteristics of participants of this study, the index that measures the prevalence of decays in adults and children (DMFT and DEFT, respectively) was at a mean of 11 for adults and 3.5 for children, which is considered very high. Despite a decline in prevalence of caries in Brazil, some studies have pointed out that the caries disease is disproportionate in reaching the economically vulnerable population. They also provide knowledge of how the epidemiology of this disease occurs in the different groups, which is fundamental for establishing preventive strategies of control [78,79,80].

Self-perception in oral health, validated as an instrument needed in treatment, is a negative self-evaluation predictor, and can also be related to a lower per capita family income and lower level of education [76, 77, 81]. Thus, self-perception can be used to identify individuals more or less vulnerable [77, 81] and who need treatment, thus corroborating its use in items of the DVS (“Do you consider it important to take care of your mouth?”; “Do you believe that oral diseases can be avoided?”; “Do you consider yourself responsible for your oral health?”; and “Do you consider it important to have all the teeth in your mouth?”.

Living conditions are strongly linked to quality of life and permeate the social inequalities present in the communities, gaining strength as one of the determining factors of the levels of health in a population [82]. Within the Brazilian territorial scenario, extensive inequality is noted as to housing, which points out the fact that the vast majority of the Brazilian population lives in houses that are not appropriate or insufficient for residential density [83], Adding this to the absence of housing conditions ideal for promoting health and well-being, as per the results found in this study, in which the variables related to living conditions within the infrastructure dimension, were consolidated as being important for defining dental vulnerability [4].

In this study, the infrastructure criteria of “presence/absence of a bathroom/electric power/running water and sewage system” remained in the DVS scale, reinforced by the literature, which indicates the lack of infrastructure services (without, at least, one of the basic services: electric power, general water mains with internal plumbing, general sewage system or septic tank, and garbage collection) compose the criteria of the household inadequacy indicators, which are determining and conditioning factors that interact in the life of the population, interfering in health promotion [84].

Among the items belonging to the infrastructure dimension, sanitation is considered by the World Health Organization as the main element of environmental health, with a significant role in the prevention and promotion of health of the population. Unfortunately, some items such as the absence of electric power, absence of sewage system and of running water are still a common reality in all Brazilian states, characterizing the precariousness of basic sanitation services [85].

The items of “healthcare services” dimension may indicate a high vulnerability of the population to oral diseases when there is no well-structured PHC service, with difficulty in scheduling visits. Access is related to the various possibilities of entering healthcare services, which would be involved with the location of the health facility, the availability and flexibility of schedules and the days the facility serves, the possibility of attending non-scheduled visits and the user's self-perception about the quality of access [86,87,88].

There is agreement on the advantages of PHC-based health systems: increasing the population's access to services is related to better performance in disease prevention and health promotion, improved health levels, reduced health inequities, and better awareness of chronic diseases among the population [89, 90].

Integrating OHT to FHS has a positive impact in increasing the access of families to curative and preventive procedures in health facilities [87]. The item "Are you accompanied by an oral health team?" is related to the access to dental services, and reflects the influence of individual and contextual determinants on the quality of oral health care. The characteristic of this access (periodicity, ease of scheduling) can be an important indicator of how the major oral health problems have been identified and followed up. It is worth noting that integrating OHT to FHS is heterogeneous throughout the national territory. Access to the dentist in certain regions results in decrease and relief of pain and suffering, not characterizing in a single way a prevention factor for oral diseases [91].

This aspect becomes even more relevant so that in the future it can assess classifications with scores, allowing professionals to use a safety and precision scale in the way of interpreting the instrument's scores.

With regard to free dental care ("Do you see the dentist without paying?"), individual income and the ability to pay for services are constantly pointed out as important factors in access to healthcare services [92]. An analysis of the use of dental services showed that in the previous 12 months, the richer individuals consulted dentists 2.8 times more than the poorest [93].

As a limitation of this study, it is important to consider the internal structure phase had its collection centered on a demographic region of the city of São Paulo, requiring expansion and testing in other scenarios. Thus, this limitation also imposes restrictions for the standardization and for the creation of the instrument's score table. So, at this point, the use of EVO should be restricted to research applications. Another aspect is the application of this scale by a large and varied number of individuals, which may make it difficult to standardize the application of the questionnaire, although barriers to avoid lack of standard have been implemented throughout this study.

Conclusion

This study shows that 15 items structured in four dimensions are sufficient to measure dental vulnerability. The DVS presents evidence of satisfactory validity, indicating its feasibility to be used by health professionals, students, and managers to plan actions and services related to oral health at PHC.

Since content validity was carried out in the national context and the internal structure concentrated in a region of Greater São Paulo, in future research the application is necessary in other regions of Brazil, expanding the evidence of validity of the internal structure, as well as studies and evidence other variables (such as the user's health indicators), and other instruments.