The number of older adults worldwide has increased to one billion in recent decades [1]. This trend poses major challenges for health systems due to the high levels of morbidity and mortality in this population due to chronic conditions (CCs), and this is a particularly significant problem in Latin America and the Caribbean (LAC), as the highest growth rate of older adults is occurring in that region [2].

Oral diseases are among the principal CCs in both the general and geriatric populations. The Global Burden of Disease Study reported that dental caries affected half of the world population in 2017 and was also the most prevalent condition assessed [3]. Some other prevalent oral diseases in the elderly are edentulism, dry mouth, periodontitis, and oral cancer, all of which have a major impact on general well-being [4]. Additionally, there is evidence that oral health influences self-esteem and socialization in the elderly population, which could be linked to a decrease in their quality of life [4]. Taking this into account, prevention of oral diseases through the utilization of oral health services is an important issue, and studies on the determinants of oral health services utilization (OHSU) in older adults are needed, as oral diseases are a frequent problem in this population.

There are socioeconomic and demographic barriers to OHSU. Several studies have found that inequalities in OHSU were associated with the level of education and welfare quintile [5]. Furthermore, the frequency of OHSU significantly varies with the area of residence [5,6]. These barriers are a common problem in countries with centralized health systems. In addition to inequalities in OHSU among older adults, the high burden of CCs among the elderly could further diminish the likelihood of their effective use of oral health services, mainly due to CC-related disabilities and out-of-pocket expenses related to the management of CCs, which may limit their access to health services [7,8].

The available literature supports the presence of associations between OHSU and CCs in high income countries. For instance, some studies have reported a link between metabolic conditions (e.g., diabetes) and OHSU, as in a systematic review conducted in 14 mainly developed Asian, European and American countries that found poor OHSU among people with diabetes, which could be linked to inadequate knowledge about oral health and poor oral health attitudes [9]. Furthermore, Okoro et al. [10] found an inverse association between the presence of depression and OHSU on a study conducted in the United States. Dental anxiety as a consequence of fear was proposed to modulate this association [11]. However, there is little evidence regarding the impact of CCs on OHSU in low-income and middle-income countries (LMIC).

Oral diseases pose significant challenges to health systems owing to their high economic burden and poor investments in oral health [12,13]. Thus, research into the determinants of OHSU requires studies to be conducted in LMIC, such as Peru, where the frequency of OHSU in older adults is low and several inequalities have been reported [14,15]. A better understanding of the associations between CCs and OHSU is needed to focus on early intervention and prevention as part of oral disease policies. Therefore, the aim of the present study was to investigate these associations in older Peruvian adults through a pooled secondary data analysis of the Demographic and Family Health Survey 2015-2017.

The flowchart for the inclusion criteria is shown in Figure 1. In total, we included 13,699 older adults. Regarding socio-demographic characteristics (Table 1), their mean age was 70.43± 8.00 years, the most common welfare quintiles were the poorest (39.2%) and the poor (18.9%), and 82.3% had health insurance. The frequency of OHSU in the previous 6 months was 18.5% (95% CI, 17.8 to 19.3). The most frequent CCs were hypertension (47.4%) and depression (15.2%).

We found statistically significant differences in the frequency of OHSU according to the presence and absence of diabetes, depression, and obesity, as well as educational level, sex, welfare quintile, place of residence, having health insurance (yes or no), and natural region. Additionally, there was a high percentage point (%p) difference in the frequency of OHSU between the extreme categories of education (higher education vs. none or elementary school: +24.8%p) and welfare quintiles (richest vs. poorest: +24.0%p). Furthermore, a relevant percentage point difference (+10.6%p) was observed between urban-dwelling and rural-dwelling participants (Table 2).

In the crude model, OHSU was associated with diabetes (OR, 1.46; 95% CI, 1.26 to 1.69) and depression (OR, 0.63; 95% CI, 0.55 to 0.72). Regarding combinations of CCs, people with diabetes and hypertension had a higher likelihood of OHSU (OR, 1.49; 95% CI, 1.25 to 1.77) than people without this combination. In the model adjusted for smoking, obesity, educational level, age, sex, welfare quintile, area of residence, insurance, and region, the only significant finding was that people with depression had a significantly lower likelihood of OHSU (OR, 0.82; 95% CI, 0.72 to 0.95) than people without depression. Finally, no association between OHSU and multimorbidity was observed in any model (Table 3).

In this study, the frequency of OHSU in older adults within 6 months prior to the survey was 18.5%. We found socioeconomic inequalities in OHSU, with particularly high percentage point differences between extreme categories of education and welfare quintiles. Specifically, a greater proportion of those who had higher education and those in the richest quintile utilized oral health services. In the adjusted model, the association between diabetes and OHSU became non-significant, but depression significantly decreased the likelihood of OHSU in older adults.

This study also revealed that OHSU was low in older adults. It has been reported that OHSU diminishes with age [22]. This could be attributed to the consequences of aging-related difficulties in communicating and functional dependence [23,24], which may compromise older adults’ social skills and mobility and may alter their patterns of healthcare utilization.

In addition, the Peruvian Basic Healthcare Plan does not include complex recuperative treatments, such as oral rehabilitation with dental prostheses (https://bit.ly/36YLHCE). This is an important non-covered treatment needed by older adults, since edentulism is highly prevalent in this population [4]. Moreover, these treatments are usually offered by private dental services, and this may explain why only 2 in 10 older adults who reported utilizing oral health services had their care covered by health insurance.

Several socio-demographic barriers limit OHSU in various populations. We found relevant inequalities regarding welfare quintile, education, and area of residence. This is supported by national and international literature [5,6,14,15]. A study using ENDES data reported that the annual frequencies of OHSU in older adults were low, and these percentages increased with welfare quintiles [14]. Other studies found educational and residence-based (urban vs. rural) inequalities in OHSU in adults in the United Kingdom and in older adults in Chile, respectively [5,6]. These findings are similar to ours. Favorable characteristics regarding individuals’ socioeconomic status are likely to improve their health status, since both education and welfare quintile promote knowledge of health care and access, respectively. Policies should improve the centralized health system and education on prevention to broaden the scope of OHSU [25,26].

Regarding CCs, the prevalence of OHSU in the previous 6 months was 24.3% in participants with diabetes. Although this was greater than the overall percentage of participants who received OHSU (18.5%) in our study, some studies of patients with diabetes have reported even higher prevalence rates of OHSU (between 40% and 73%) [27-29]. However, those studies evaluated dental visits during the previous 12 months, unlike our approach, which involved a 6-month interval. Other studies in older Peruvian adults have also reported low prevalence rates when evaluating OHSU over a 12-month (21.7%) or a 6-month period (24.9%) in 2014 and 2018, respectively [15,30]. Despite the pathophysiological link between oral health and diabetes [31], which involves a greater need for OHSU, in our study this association became non-significant after controlling for socioeconomic confounders. Most likely, a low knowledge of oral health, which may be determined by education, impacted the degree to which persons with diabetes were interested in OHSU. The ENDES does not report when diabetes was diagnosed; however, it would have been useful to control for this factor when analyzing this association, since a longer duration of diabetes is associated with more frequent dental manifestations. Likewise, welfare quintile—as a proxy for socioeconomic status—influences health access, and, given the fact that oral health services are not available to everyone, people in the lower welfare quintile would be likely not to receive oral health services despite having diabetes. Likewise, a similar association was described in the study of Wade et al. [32] in the United Kingdom, in which the probability of OHSU was lower in adults with 1 CC and a low educational level. Additionally, a study of 41,220 participants in Germany found that lower socioeconomic status was associated with a lower likelihood of OHSU [33]. It is important to note that some studies have reported that patients with diabetes who were better informed about oral manifestations were more likely to have good attitudes towards oral care [25]. These results show that despite the need for OHSU, education and other socioeconomic proxies have a powerful influence on the utilization of these services.

We found a high frequency of non-OHSU (86.8%) in the previous 6 months among older adults with depression. Two studies reported a lower frequency of non-OHSU (31.6-43.7%) in adults with depression in the last year [10,34]. However, aging influences depression, and it is therefore important to keep in mind that one of those studies did not include older adults, and the prevalence of older adults in the other study was 15.9%, whereas our study only assessed older adults. Moreover, OHSU was considered in the last year or 2 years in these studies, whereas we considered the last 6 months. Lengthening the study period allows more cases of OHSU to be found, with a subsequent reduction in the intensity of the association between non-OHSU and depression. The distribution of age in the study population and the method used to measure OHSU are relevant characteristics for understanding differences among studies in the frequency of OHSU in people with depression.

Furthermore, in the adjusted model, people with depression had a lower likelihood of OHSU. Several papers have suggested possible mechanisms to explain this association. For example, symptoms of geriatric depression, such as chronic pain [23,35], functional dependence [23], and fatigue, in addition to geriatric frailty, can limit the degree to which they notify others of their discomfort, leading them to avoid social situations and consequently causing reductions in their frequency of and interest in healthcare utilization. This situation may be intensified by changes in communication ability in the elderly. For example, 75% of 12,769 older adults with Medicare (United Sates) had problems with communication (listening, writing, using the phone) [33,36]. Moreover, the highest regional growth rate of older adults (in the 2015-2020 period) can be found in LAC, and since the majority of these countries are LMIC [2], they may not be able to address communication issues in this population, thereby reducing the effectiveness of healthcare services.

Another interesting explanation for the low utilization of oral health services in people with depression is their fear of visiting the dentist. People with depression may be afraid of threatening stimuli in dental situations, and this may manifest as devastating feelings of phobia and trepidation. Bernson et al. [11] reported that dental fear, anxiety, and depression were significantly higher in people with irregular dental care than in people with regular OHSU. Older adults generally cannot attend health centers on their own and must almost always be accompanied due to communication and mobility difficulties. Another study showed that the frequency of depression and anxiety was higher in patients with high dental fear than in those who had mild fear [37]. Psychological conditions are linked with negative emotions and socialization problems, which can intensify dental fear.

Due to the high frequency of OHSU in older adults with CCs, we hypothesized that multimorbidity could be associated with OHSU; however, this was not the case in our study. In contrast to our findings, Wade et al. [32] found that multimorbidity increased the likelihood of OHSU in adults in the United Kingdom. While the objective of their study was similar to ours, they included more CCs (12 vs. 4), which increased the frequency of multimorbidity in participants with OHSU (40.0% vs. 32.9%), meaning that the association could have been overestimated. Although people with multimorbidity use health services more than people without multimorbidity [38], it is erroneous to consider that all CCs have the same impact on OHSU. This is because the oral complications or manifestations of CCs vary in prevalence, extent, and severity [39,40], which may influence health-seeking behavior, and therefore, OHSU.

The results of this study document an association between depression and OHSU in older adults, showing a negative relationship between these 2 factors. Improved policies are needed to ensure a national scope of care for the minimum dental requirements of older adults, as well as policies that make this care more available for specific groups, such as people with mental health problems. New strategies must be focused on overcoming barriers, which are mainly related to education, welfare quintile, insurance and area of residence.

Regarding the limitations of this study, this was a secondary analysis of a survey; therefore, the data could have been affected by inaccuracy or memory bias. Although the investigators received previous training, there could have been errors in data collection. However, as the ENDES follows the DHS model, there is a certain methodological strength in the data collection. Consequently, the bias may have been lower than would otherwise be expected. The OHSU did not distinguish between recuperative and preventive interventions, nor were more frequent CCs such as osteoarthritis reported in older adults, which would be interesting to analyze. Furthermore, there was a 6-month window period for this survey, which could have excluded people who had utilized oral health services in the previous months. Despite these limitations, this survey is national in scope with a high level of regional representativeness, and the results are important for new national policy proposals regarding oral health in older adults.

In conclusion, the prevalence of OHSU was found to be low in older Peruvian adults. Socioeconomic variables, such as education, welfare quintile, having health insurance, and area of residence, were determinants of variation in OHSU. Regarding CCs, we found that the association between diabetes and OHSU depended on demographic and socioeconomic variables, while depression reduced the likelihood of OHSU independently of these variables. Although further research is needed, our results may be useful for the development of policies aimed at increasing OHSU in older adults with CCs, especially in those with depression.