The data used come from the annual component of the 2013-2014 Canadian Community Health Survey (CCHS), conducted by Statistics Canada. . This is a broad representative survey of the Canadian population and at the provincial level.
The CCHS includes both a set of core mandatory questions that are completed by respondents in all provinces and a set of optional components completed at the discretion of each province. We based our analysis on the full national sample with non-missing values on all variables used and the adult population over 17 years of age. We first estimated models for the likelihood of using dental care using CCHS data for all major Canadian regions. For our purposes, respondents living in Ontario (about one-third of the population) were asked questions about unmet dental care needs, which we used to study access to dental care for that province only, and we estimated models for the likelihood of reporting unmet need. need for dental care with the richer set of independent variables on dental care that was only available for Ontario (CCHS dental care inclusion modules 1 and 2). Details of the selection process for the datasets used are reported in additional online resources (Supplementary File 1).
Our two dependent variables measured use of dental care and lack of access to dental care. The measure of dental care utilization was based on a standard question asking whether people had seen a dentist, dental hygienist, or orthodontist in the 12 months prior to the interview (consultation meaning either seen or talked to).Footnote 1 To study access, we used a question asked of respondents living in Ontario (and the Northwest Territories) only, regarding not seeing a dentist in the past 3 years. Respondents who reported no visits in the past 3 years were categorized as having unmet access to dental care, assuming that everyone should see a dentist at least once every 3 years. We also used a more subjective variable in a sensitivity analysis: those who had not reported any visit in 3 years were asked the reason, and we created an unmet access variable for financial reasons by assigning the value 1 to respondents indicating the reason was because of cost. Finally, we used a question posed to Ontario respondents only to perform a sensitivity analysis on our analysis of dental care utilization (first analysis):It has been previously reported that you have “seen” or “talked” to a dentist in the past 12 months. Have you actually visited one? the variable on use in the past 12 months being restricted to those who had visited one.
Need variables included age, gender, poor self-rated health, poor self-rated mental health, and low perceived life satisfaction. Following previous studies we tested the inclusion of additional chronic health indicators, including a dummy set for the main chronic diseases reported by respondents (rheumatic, cardiovascular, respiratory diseases, diabetes, cancer, digestive diseases, back problems and scoliosis, anxiety and other chronic diseases).Footnote 2 Subsequently, we retained the only chronic condition variable that had a significant impact in all our models, the fact of having been diagnosed with diabetes. Unfortunately, information on oral health (a potentially important determinant of the need for dental care) was only available for respondents living in Ontario. We performed a sensitivity analysis on this subpopulation, including information on oral health, and concluded that its inclusion did not change the statistical significance or the values of the other estimated effects. For the analyzes carried out with the respondents from Ontario, we used the following variable for oral health: we constructed an index of oral health status by Cronbach’s alpha, comprising a set of 14 questions on the oral health conditions.Footnote 3 We then used a variable that takes the value 1 if the individual belongs to the lower tertile of the distribution of this index, declaring a poor state of dental health, and zero otherwise.
Variables without need
We divided income by the square root of household size and then took the logarithm of that equalized income. Food insecurity (FI) is measured by Statistics Canada with a household food insecurity status indicator, based on a set of 18 questions, and describes the household’s food security status over the past 12 months . It captures three types of situations: 1–Food security: no sign of difficulty in accessing food linked to income; 2–Moderate food insecurity: sign of compromised quality and/or quantity of food consumed; 3–Severe food insecurity: sign of reduced food intake and disturbed eating behavior. This variable is taken from Health Canada’s food security status model. FI is measured in the United States and Canada using a standard questionnaire. We used data from the group food security module of the CCHS survey. This is based on the definition of FI as “uncertainty and inadequacy of food availability and access that is limited by resource constraints, and worry or anxiety and hunger that may result” (Wunderlich and Norwood, p.49) . Households classified as food secure did not report any problems. Households classified as moderately food insecure had problems with the quality and/or quantity of food consumed by adults and/or children, while severe food insecurity had greater trade-offs, including reduced food intake in adults and/or children due to a lack of money. for food (see Statistics Canada ; Tarasuk ).Footnote 4 We therefore used two dummy variables if the individuals lived respectively in a family with severe or moderate food insecurity, and zero otherwise. Housing status in the survey was captured by a housing tenure question, where individuals were asked whether they lived in rented accommodation or owned. We used the dummy variable for rented accommodation as an indicator of housing insecurity, as opposed to housing security with home ownership. There was no possibility of distinguishing between those who owned their homes with or without debt. In our estimates, we controlled for a range of other non-needs variables, such as marital status, immigrant status, race/ethnicity (Indigenous, non-white), smoking status ( daily smoking), obesity, regular alcohol consumption and rural living. areas. The provinces in which individuals live were grouped into six regions, and a dummy variable for each was created: (1) Atlantic, which includes Nova Scotia, New Brunswick, Prince Edward Island and Newfoundland and Labrador; (2) Quebec; (3) Prairies, which includes Alberta, Saskatchewan and Manitoba; (4) Territories, which includes Yukon, Northwest Territories and Nunavut; (5) British Columbia; and (6) Ontario.
We added a variable related to dental hygiene lifestyle. In the special module available only for Ontario, people were also asked how often they brush their teeth (“How often do you brush your teeth?”). We used a dummy variable what youtakes the value 1 if it was at least twice a day, and zero otherwise. Finally, we used another dummy variable, taking the value of 1 if the individuals were covered by dental insurance, and zero otherwise. Our reference person for models on the use of dental visits was a middle-income white, married (or common-law) local citizen, middle-aged woman (46-55), living in Ontario, not actively working as an employee or self-employed. i.e. being unemployed, doing household chores, being a student, on welfare or in other non-working conditions, having a secondary education level, having at least one medical condition/ acceptable mental health/life satisfaction, not affected by diabetes, not obese, not currently smoking or heavy drinking, not physically active, not food insecure, not having to pay rent for a house, living in an urban area. For models on unmet access to dental visits, the reference person had the same characteristics as above and, in addition, had at least fair or good oral health, no dental insurance and no did not brush his teeth at least twice a year. daytime. The descriptive statistics of the main dependent and independent variables used are presented in Table 1.