Food insecurity, home ownership and income-related equity in dental care use and access: the case of Canada

There is some evidence of socio-economic inequalities in oral health [1, 2]. Pro-rich socio-economic inequalities and income-related inequities in the utilization of some health care services in Canada, particularly dental care, have been documented [3‐10]. Also, during and after the global financial crisis of 2008–9, food insecurity and housing problems have increased in Canada as well as in the US and Europe [11, 12]. There are good reasons to believe that food insecurity and home ownership may affect the extent of social inequalities and income-related inequity in dental care use in Canada. In this study, we propose to enrich the description of socio-economic status (SES) in assessing equity in access and utilization of dental care by adding these two measures of household economic and financial insecurity as non-need independent variables beside income. Annual income provides much information on what commodities a household can afford and, as a result, whether the household can afford dental care services without jeopardizing its other consumption. We posit that it does not, however, capture all relevant information, and we propose to enrich the description of SES by adding these two household characteristics. Food insecurity is a measure of how likely it is that a household cannot afford the required quantity and quality of food [11]. A general definition of household food insecurity is given by Anderson [13] and refers to having a limited or uncertain availability of nutritionally adequate and safe food, or to having to acquire foods in socially unacceptable ways. This is a measure of purchasing power, and we posit that purchasing power may vary across households with similar income levels, especially since the big recession of 2008. The second characteristic is home ownership: here, we hypothesize that two households with the same income may have different purchasing power and situations relative to uncertainty depending on whether they own their house, especially in Canadian urban centres where housing prices have increased dramatically [14, 15]. As a result, home ownership status may affect the affordability of dental care services, conditional on income, and we want to test this is the case. There is evidence for both Canada and the US that both food insecurity [16‐25] and housing insecurity [26‐29] have a negative impact on individual physical and mental health. Duncan and Bonner [30] for Canada find a detrimental effect of food insecurity on oral health. Access to healthy food is considered a factor that affects oral health [31, 32]. There is conflicting evidence on the link between food insecurity and health care use. Tarasuk et al. find an effect for Canada for ambulatory care [33] and for mental health care [25]. Kushel et al. [26] find an effect for the US and both ambulatory and emergency care. However, Bhargava and Lee [34] find no effect among the 50+ in Georgia, US. On dental care in particular, Allin et al. [5] find that, in Europe and the US, inequalities in dental care use are more pronounced across wealth than income quintiles. Muirhead et al. [35] show that food insecurity has a powerful impact on dental care use among Ontario working poor: 40% of those who are food insecure report seeing a dentist only when in pain, versus 18% of those who are food secure; 52% of those who are food insecure reported unmet dental care need, versus 24% of those who are food secure. Previous work on unmet needs for dental care in Canada found evidence that individuals with low income and without dental insurance coverage had the highest probability of reporting financial barriers to dental care and that reported unmet needs due to costs were associated with lower dental visit frequency and poorer oral health outcomes after controlling for the effects of income and insurance coverage [36]. However, the impact of both the lack of house ownership and food insecurity on income-related inequities in dental care use and unmet access to dental care has never been studied. We will add to the literature on income-related inequities by introducing these two variables among other socio-economic (non-need) determinants and by estimating their contribution on horizontal equity in dental care use. We will also add a rich set of controls describing chronic conditions, testing their role on dental care utilization. It has been shown for other countries like Italy that equity in the use of and unmet needs for health care services may vary not only by region [37] but also by type of chronic disease [38]. There is evidence for Canada that the use of health care services varies with the type of chronic conditions [4] and that being affected by diabetes tends to be associated with lower access to dental care [10].

However, evidence is needed on the role of food insecurity and home ownership in income-related inequities in access to dental care in Canada.

The data used were from the 2013–2014 Canadian Community Health Survey (CCHS) annual component, conducted by Statistics Canada [39]. This is a large survey representative of the Canadian population and at the provincial level.

CCHS includes both a mandatory core set of questions that are completed by respondents in all provinces and a set of optional components completed at the discretion of individual provinces. We based our analysis on the full national sample with non-missing values on all the variables used and the adult population aged more than 17 years old. We firstly estimated models for the probability of using dental care by using CCHS data for all Canadian main regions. For our purposes, respondents living in Ontario (approximately one-third of the population) were asked questions on unmet need for dental care, which we used to study access to dental care for that province only and we estimated models for the probability of reporting unmet need for dental care with the richer set of independent variables on dental care that was only available for Ontario (dental care inclusion modules 1 and 2 of the CCHS). The details of the selection process of the datasets used are reported in additional online resources (Additional file 1).

According to the principle of horizontal equity in health care, access should depend only on need, while socio-economic factors unrelated to need should not influence utilization [41]. We tested for income-related equity in health care use, applying the indirect standardization by regression method [41‐45]. The horizontal inequity index (HI) is defined as the difference between income-related inequality in the observed health care use (CM) and income-related inequality in “need-expected” use (CN). The need-expected use is obtained by setting at their sample mean all non-need variables. The concentration index CM is positive (negative) whenever better-off individuals (worse-off) use dental care more than worse-off individuals (better-off). If the need distribution is favouring those who are worse off (better off), then we observe that the value of CN is negative (positive). There is no inequity in access to care when CM equals CN or the difference is not statistically significant. A positive (negative) value of the inequity implies inequity favouring those who are better off (worse off). Conversely, when the dependent variable measures unmet access to dental care, a negative (positive) value of HI implies inequity favouring those who are better off (worse off). An intuitive interpretation of the HI results can be obtained by multiplying HI by 75; for example, a HI value of 0.2 implies that equity can be achieved by redistributing 15% (0.2 × 75) of care from the rich to the poor [41]. For more details on the methodology see the additional resource (Additional file 2). In order to estimate CN, we estimated two set of logit models for: 1–the probability of having consulted a dentist or an orthodontist in the last 12 months in Canada; 2–the probability of having visited a dentist or an orthodontist in the last 12 months and of reporting lack of access to dental visits in the last 3 years in Ontario. The model for the lack of access to dental visits in Ontario was also estimated for the subset of individuals that indicated the main reason for not having visited a dentist in the last 3 years to be dental care costs. In all models we distinguished independent variables between need and non-need variables (Table 1). Non-need variables were set to their sample mean for obtaining CN [41, 45]. We estimated the income-related HI index for 1) the probability of using dental care services (all Canada and provinces) and 2) for the probability of unmet access to dental visits in the last 3 years (Ontario). All estimates were derived from weighted models using the population weights provided by Statistics Canada with the survey data. The estimated standard errors were based on the bootstrap method using weights provided by Statistics Canada.

The proportion of Canadians visiting a dentist in 2013–2014 was approximately 67%; the proportion of those who reported they did not access a dental visit in the last 3 years—our measure for unmet access to dental care¹—in Ontario and Northwest Territories was 16% (Table 1), and 14% in Ontario alone.⁵ Of all individuals in Canada, 6% were living in food insecurity (4% moderate and 2% severe) (Table 1). Those living in a rented house were 27% of the total population, but this proportion more than doubles among the moderately food insecure (63%) and is almost three times higher among the severely food insecure (75%). Those who reported dental visits in the past 12 months were slightly less likely to be moderately (3.9%) or severely (1.6%) food insecure than the general population, and only 19% were tenants. These proportions were much higher among individuals reporting not having seen a dentist in the last 3 years in Ontario, of which approximately 17% were moderately (11%) or severely (6%) food insecure and 44% were tenants. Among those who reported unmet access, 30% were tenants and had food insecurity (either moderate or severe), i.e., five times the proportion of those who did not report problems in access (6%).

Overall, individuals who were better off were more likely to use dental care than the poor and less likely to report unmet access to dental visits during the 3 years preceding the interview. Tables 2 and 3 show the estimated odds ratios for all the logistic models. As expected, the probability of visiting a dentist or an orthodontist in Canada and of reporting problems in access to dental care is overall significantly associated with health care needs. With respect to the reference individual, being male, younger or elderly, being obese, and having diabetes are all factors associated with lower odds of use and higher odds of unmet access to dental care. Higher income is associated with higher odds of using dental care and conversely of unmet access to dental care, as high education and being employed. People with a high level of education and living in British Columbia were also more likely to access dental care. Ethnicity, marital status, and lifestyle all have associations with dental care access: being non-white or aboriginal, being a widow, being obese, smoking, and regular alcohol use are all associated with lower odds of using dental care. Engaging in regular physical exercise is associated with higher odds of visiting a dentist.

The results show that food insecurity is an important non-need factor of both equity in the use of and access to dental care. In particular, food insecurity is associated with lower odds of using dental care (Table 2) and higher odds of reporting lack of access to dental care (Table 3). Importantly, another component of financial insecurity related to housing, as measured by having to pay rent and not owning a house, is an important non-need factor in both horizontal equity in the use (Table 2) and unmet access to dental care (Table 3). Overall, the probability of visiting a dentist would be higher if people were not living in high food insecurity and/or owned their dwelling (Fig. 1). There is pro-rich inequity in the probability of visiting a dentist or an orthodontist (Table 2). Inequities vary across jurisdictions. Atlantic (HI = 0.14; P = 0.008; CI:0.000 0.129) and Quebec (HI = 0.13; P = 0.000; CI:0.113 0.148) show the highest inequity in the probability of dental care, Ontario (HI = 0.11; P = 0.000; CI:0.103 0.123) and the Territories (HI = 0.12; P = 0.000; CI = 0.071 0.152) are similar to the national average (HI = 0.11; P = 0.000; CI:0.105 0.117), whereas British Columbia (HI = 0.09; P = 0.000; CI: 0.077 0.11) and the Prairies (HI = 0.10; P = 0.000; CI: 0.076 0.108) show values below the national average (Fig. 2, Table 2). In comparison with previous articles based on data for the period 2000–2005, dental care is still distributed in favour of the richest [5, 7, 9]. We obtained almost identical results by using either the more general question, “Have you seen a dentist or orthodontist”, available at the national level for all regions, or the more specific question, “You reported that you have seen a dentist or orthodontist in the last 12 months: have you actually visited one?” (Table 3). This confirms the robustness of results obtained at the national and regional levels without the extended set of variables on dental care. There are high inequities in unmet access to dental visits in Ontario (HI = -0.39, P = 0.000, CI:-0.4276 0.3443); again, high food insecurity and housing insecurity appear autonomous non-need factors contributing to the probability of reporting problems in access to dental care (Table 3). Our results show that the HI for unmet access does not change substantially using the richer set of variables from the special modules variables on dental care inclusion in Ontario if the reason for unmet visits is due only to costs (HI-0.44, P = 0.000, CI: − 0.5209 0.3623) (Table 3). Estimates for Ontario show that high and moderate food insecurity are associated with higher odds of reporting unmet access to dental care and lower odds of using dental care. Differently from the countrywide estimates, the odds of high food insecurity are higher than the odds of moderate food insecurity for both use of and unmet access to dental care. Similar to the nationwide estimates, living in a rented house is associated with lower odds of using dental care and higher odds of having problems in access (Table 3).

We can see from these results that not having food security and living in a rented house are associated with lower probability of using dental care, everything else being the same. These two variables also affect access to dental care in Ontario. Like in other countries such as Italy, we found that inequities in utilization are smaller than those in unmet access to dental care [37]. In the case of Ontario, inequities in access are three times higher than inequities in use of dental care. Estimates obtained by using the extended set of variables available for Ontario are similar to those obtained with the common set of variables available at the country level. This reinforces the results obtained with the core set of variables available nationwide. Our data are cross-sectional so we cannot infer causality. However, our evidence for Canada reinforces the view that reducing food insecurity in Europe and North America could counterbalance the negative effect of food insecurity on health care use [33, 46, 47]. Results are in line with previous evidence for Ontario showing that public spending on housing can offset the relationship between rising unemployment and food insecurity [47]. As a further test, we performed a decomposition analysis of the HI index for the probability of using dental care.⁶ This confirmed the results that housing and food insecurity are relevant determinants, together accounting for the largest contribution after income on the HI index (see Additional file 3).⁷

Estimates of the HI (horizontal inequity) index for the probability of using dental care show that, overall, inequities in dental care use in Canada are persisting. This work adds to the literature by also estimating inequities in unmet access to dental care in selected areas of Canada for which data were made available (Ontario and NW Territories). As expected, in these areas inequities in access are much higher than inequities in use. Therefore, it seems important for Canada to systematically report on indicators not only on the use of dental care but also on unmet access to dental care. The methodology applied aimed at describing the contribution of other non-need factors to horizontal inequity than the traditionally considered such as income, education etc. We newly showed that both food insecurity and home ownership are non-need factors contributing to income-related inequity in dental care access and use. This study is based on cross-sectional data, and therefore inference on causality is limited. We cannot interpret this as a causal relationship from food and housing insecurity to dental care use, as these two variables can be influenced by unobservable characteristics that can also affect dental care use. We could not fully explore the links among insurance coverage, food insecurity, and home ownership due to data limitations at the national level. Future work will be targeted at analysing in detail such links by allowing the confounders and mediators of the relationship between each of the covariates and dental services to vary. Moreover, further work is required to see if and how extended insurance coverage for dental care could at least partially offset the difficulty of living in food insecurity and without home ownership in terms of improving access to dental care. Despite its limitations, this study favours the argument that it is important to tackle the food and housing insecurity of households in order to reduce existing inequities in access to and use of dental care, thus improving the overall equity performance of Canada’s health care system. There is a need for interventions aiming at improving equity in access to dental care, such as increasing public dental care coverage for fragile groups. Our study suggests that redistributing income may not be enough to reduce inequities in dental care use and access to dental care. What matters is to also act on other dimensions of purchasing power, such as housing tenure status and food security.