Oral Health–Related Quality of Life Among Publicly Insured Mental Health Service Outpatients With Serious Mental Illness

Individuals with serious mental illness, such as schizophrenia or bipolar disorder, tend to have worse oral health than individuals without these psychiatric conditions, including higher rates of untreated caries and temporomandibular joint disorder, increased tooth decay and tooth loss, greater need for tooth extractions, and worse periodontal health (1–6). Concerns about their patients’ oral health rank high among providers who treat persons with serious mental illness (7), and there is increasing acknowledgment that improving oral health can improve mental health (8).

The causes of the high prevalence of poor oral health among individuals with serious mental illness are unclear. Higher than average poverty rates (9) likely contribute, given that low socioeconomic status is associated with poor oral health (10). Individuals with serious mental illness may be more exposed to barriers to dental service use, such as high dental costs and limited dental insurance coverage, resulting in poor oral health (11, 12). Dental anxiety, a form of apprehension or even phobia about dental procedures, may be elevated among individuals with psychiatric disorders and potentially hinder dental care (13, 14). Limited English proficiency (LEP) may also hamper access to dental services among patient subpopulations (15). However, even when persons with serious mental illness attend dental appointments at similar rates as individuals without these conditions, they are more likely to have unmet dental needs (16), exposing the limitations of the dental services available to them and the role of factors other than service use.

Other factors potentially associated with poor oral health among individuals with serious mental illness include xerostomia (oral dryness), due to the anticholinergic effect of psychiatric medications and associated with caries and periodontal disease (17–19); smoking, common among individuals with psychiatric disorders and associated with higher risk of periodontal disease and oral cancers (20, 21); and self-care limitations and poor dietary practices associated with psychiatric diagnoses, especially schizophrenia and related disorders (4, 22–24).

To help identify modifiable targets for oral health improvement initiatives among individuals with serious mental illness, we examined potential factors associated with poor oral health in a public mental health outpatient program serving a largely low-income population primarily of individuals from racial-ethnic minority groups, underserved groups at risk of poor oral health (10, 19, 25) and low dental service use (26, 27). First, we examined the association of past-year unmet dental need with past-year dental service use, financial and nonfinancial barriers, dental anxiety, and LEP. Next, after controlling for xerostomia, smoking status, schizophrenia spectrum disorder, and other factors, we examined the association of past-year unmet dental need with low oral health quality of life (OHQoL), a multidimensional subjective assessment associated with health-related quality of life and general well-being (28, 29) that encapsulates functional (e.g., eating) and psychosocial (e.g., self-esteem and social functioning) dimensions (30–32).

Methods

Sample

Participants were recruited by using convenience sampling from a public mental health service treating patients with serious mental illness in upper Manhattan in June–September 2014. Patients self-selected for participation by approaching study staff at research tables in the clinic or by completing a sign-up sheet. Study staff screened interested patients for the following eligibility criteria: English or Spanish fluency, absence of dementia on a screening instrument for patients over age 64 (33, 34), and capacity to consent (35). Bilingual research staff conducted interviews in the participant’s choice of language (English or Spanish). Published Spanish translations were used, except for the xerostomia scale, which was forward-translated with bilingual committee review. The Institutional Review Board of the New York State Psychiatric Institute approved this study.

We assessed 189 patients for participation, excluding 39 patients (eligible [incomplete interview, N=1; declined participation, N=9; missed screening, N=15] or ineligible, N=14), for a total of 150 participants.

Measures

OHQoL.

The 14-item Oral Health Impact Profile (OHIP-14) (36, 37) assesses OHQoL (38, 39). Items rate the frequency of the impact of oral health problems in the past 12 months on a scale from 0 (never) to 4 (very often), with higher scores indicating worse OHQoL. The OHIP-14 is reliable in diverse settings and populations, including Latinos (40). Internal reliability in our sample was high (Cronbach’s α=.93). We defined low OHQoL as any OHIP-14 item occurring at least fairly often (37, 41).

Unmet need for dental care.

Past-year unmet need for dental care was based on a binary item from the 2013 National Health and Nutrition Examination Survey (NHANES) section on oral health (42) that asked whether the participant needed dental care in the past year but was unable to obtain all the care that was needed.

Dental visit.

Past-year dental visit was based on patients’ self-reported last visit to a dental care provider (dental hygienist, dentist, orthodontist, oral surgeon, or other dental specialist) during the past year.

Perceived barriers to seeking or receiving needed dental care.

Because cost is the most common barrier to dental care (11, 43), we examined financial and nonfinancial barriers separately using 2013 NHANES oral health items (42). Financial barriers (three items) were lack of dental insurance, inability to pay for dental services, and concern about treatment cost. Nonfinancial barriers (11 items) were lack of transportation, excessive wait for appointment, inability to obtain needed appointments, inability to take time off work, language problems, inconvenient clinic hours, patients without someone to accompany them to the appointment, being excessively busy, fear of visit due to immigration status, not knowing where to seek services, and not thinking problem was serious or expecting it to go away. Both types of barriers were analyzed dichotomously (any/none).

Xerostomia.

The five-item Summated Xerostomia Inventory (SXI) assesses the frequency of dry mouth symptoms on a 3-point scale (1, never; 2, occasionally; and 3, often). Possible scores range from 5 to 15, with higher scores indicating greater severity. This measure performs reliably, with scores associated significantly with a gold-standard measure of oral dryness (44). The SXI had adequate reliability in our sample (Cronbach’s α=.75). SXI score was centered on the sample mean score of 7.72.

Dental anxiety.

The five-item Modified Dental Anxiety Scale (MDAS) (45) measures participants’ expected anxiety at various time points before a dental visit, from the day before to the moment of treatment, on a 5-point scale, from 1, not at all anxious, to 5, extremely anxious. The scale has high reliability (Cronbach’s α=.89) (46), with good reliability in our sample (Cronbach’s α=.82). Summed scores (possible range, 5–25) were analyzed dichotomously (high anxiety, ≥19, and low anxiety, <19) (47, 48).

Psychiatric diagnosis.

Analyses of self-reported psychiatric diagnoses compared outcomes for schizophrenia spectrum diagnoses (schizophrenia or schizoaffective disorder) versus other diagnoses.

Data Analysis

We examined the bivariate associations of past-year unmet dental need with past-year dental visit, barriers in seeking or receiving care, age (5-year units), gender, and LEP (speaking English less than very well) (49). We then examined their adjusted associations using a multivariable binomial logistic regression model. We also examined bivariate associations of low OHQoL with past-year dental visit, xerostomia, age (5-year units), current smoker (smoking ≥5 cigarette packs over the lifespan and currently smoking every day or some days) (50), and schizophrenia spectrum diagnosis; adjusted associations were estimated with a multivariable binomial logistic regression model. We obtained the adjusted risk ratio (ARR) (the ratio of the mean predicted probabilities) and the adjusted risk difference (ARD) (the difference in the mean predicted probabilities or the excess disease burden associated with a factor) using the adjrr package in Stata SE 12.1 (51). This follows recommendations to report relative and absolute measures of association which can provide comprehensive information useful to patients and providers (51, 52). All significance tests were two-sided with α=.05.

Results

Participants (N=150) had a mean±SD age of 50.77±14.98 years and were predominantly female (63%) (Table 1). Most participants were Hispanics (78%), were born in the Dominican Republic (51%), spoke Spanish at home (72%), and had LEP (68%). Most had a high school education or less (73%) and an annual household income under $20,000 (89%). Almost all (94%) were unemployed, disabled, or otherwise out of the labor force. About half of participants reported a lifetime schizophrenia spectrum disorder (53%), and 21% were current smokers. The mean SXI score (7.72) indicated a level of xerostomia in the lower half of the range, and 27% of participants had high dental anxiety. All participants were covered by Medicaid (58%), Medicare (14%), or both (28%).

Low OHQoL was reported by 54% of participants, a past-year dental visit by 61%, and past-year unmet need for dental care by 39% (Table 1). Almost half of participants reported a financial barrier to seeking/receiving needed dental care (47%), with cost concerns reported most commonly (45%). Nonfinancial barriers to seeking/receiving needed dental care were reported by 57% of participants; the most common barriers were thinking that the problem was not serious or expecting it to go away (28%), long wait for appointment (28%), and inability to get an appointment when needed (24%). In post hoc analyses, 44% (N=57) of participants with Medicaid coverage reported any financial barrier compared with 67% (N=14) of participants with Medicare alone (χ²=3.66, df=1, p=0.056).

We examined correlates of past-year unmet dental need initially with unadjusted associations. Nearly 58% (N=33 of 57) of patients who reported past-year unmet need for dental care had a past-year dental visit; this association was nonsignificant. Among participants experiencing any barriers, the proportions reporting an unmet dental need were significantly greater than the proportions without such a barrier (63% with financial barrier versus 16% without, χ²=35.63, df=1, p<0.001; 55% with nonfinancial barrier versus 16% without, χ²=23.74, df=1, p<0.001) (Table 2).

In the adjusted model, participants with financial barriers (ARR=3.16) and nonfinancial barriers (ARR=2.18) remained significantly more likely to have past-year unmet dental need than participants without these barriers. In terms of absolute differences, participants reporting any financial or nonfinancial barriers had higher risk of past-year unmet dental need by 40 and 27 percentage points, respectively, than those without such barriers. Each 5-year increment in age was associated with greater risk of unmet need in terms of absolute difference (ARD=1.02, 95% confidence interval [CI]=0.00–0.03) but was nonsignificant in terms of relative difference. The associations of gender, high dental anxiety, and LEP with past year unmet need were nonsignificant (Table 2).

A greater proportion of participants with past-year unmet need had low OHQoL than those without unmet need (67% versus 46%; χ²=5.94, df=1, p=0.01) (Table 3). However, the proportion of participants with low OHQoL did not differ by whether participants had a past-year dental visit (54% [N=49 of 91]) versus without a visit (54% [N=32 of 59]).

After control for the relevant factors listed in Table 3, past-year unmet dental need remained significantly associated with increased risk of low OHQoL (ARR=1.31, 95% CI=1.00–1.70), with an absolute difference of 15 percentage points (95% CI=–0.00 to 0.30) (Table 3). Xerostomia score (ARR=1.20) and schizophrenia spectrum disorder (ARR=1.33) were also associated with higher risk of low OHQoL. A 1-point increase in the mean xerostomia score increased the adjusted risk of low OHQoL by 11 percentage points, and having a schizophrenia spectrum disorder increased the adjusted risk of low OHQoL by 15 percentage points. Current smoker status was not significantly associated with risk for low OHQoL in both relative and absolute differences. The association between age and adjusted relative risk of OHQoL was nonsignificant, but each 5-year increment in age significantly increased the adjusted absolute risk of low OHQoL by 2 percentage points.

Discussion

Our exploratory study of the oral health of underserved patients with serious mental illness in a public outpatient mental health program found a high prevalence of unmet dental need and of low OHQoL, despite a high rate of dental visits. Participants with a financial or nonfinancial barrier were more likely to have past-year unmet dental need. Significant independent associations with low OHQoL were found for past-year unmet dental need, xerostomia, and schizophrenia-spectrum disorder, but past-year dental visit was not associated with low OHQoL.

Over half of the participants reported low OHQoL (54%), much higher than the prevalence of 15% found in a U.S. population study that used a version of the OHIP that is psychometrically comparable to the version used in our study (37). Our finding is consistent with findings from the United Kingdom of elevated low OHQoL among people with severe mental illness compared with a national sample (53). This finding is notable given the high rate of past-year dental visit (61%) in our sample, greater than rates of 31% for New York State Medicaid enrollees in 2013–2015 (Behavioral Risk Factor Surveillance System data) (54) and 42% for U.S. adults in 2014 (Medical Expenditure Panel Survey [MEPS] data) (55). It indicates elevated vulnerability in our sample that is likely associated with exposure to intersecting clinical and demographic risk factors (10, 26). It also warrants further research on strategies to improve oral health among individuals with psychiatric disorders, given that this may lead to improved mental health outcomes (8).

Unmet dental need may partly explain why we observed elevated low OHQoL reporting. Over one-third of participants reported any past-year unmet dental need, compared with 10% of participants with serious mental illness in the nationally representative MEPS (16). Unmet need remained independently associated with low OHQoL in our sample, even after adjusting for clinical factors.

Our finding that financial barriers remain a major hurdle to reducing unmet need in a publicly insured sample is consistent with national data and supports the expansion of covered dental services and benefits. Nationally, cost is the top barrier to a past-year dentist visit regardless of age, income, and type of dental insurance coverage (56). Exploratory results suggest that financial barriers to dental care may be more common with Medicare than with Medicaid but the difference was nonsignificant. Further examination is needed to understand variations in unmet dental need associated with different forms of public insurance. Medicaid dental coverage varies widely by state; in November 2018 only 19 states, including New York, provided “extensive” coverage, defined as including more than 100 covered procedures and having a spending cap that exceeds $1,000 per person per year (57). At the time of data collection, New York’s Medicaid coverage excluded root canals, immediate dentures, and most periodontal surgeries (58). Recent changes in New York’s Medicaid program cover prosthodontics (full/partial dentures) and some dental implants (e.g., molar root canals) (58) but require prior approval and substantial documentation. Financial barriers may be worse for those with original Medicare or traditional fee-for-service Medicare, which excludes routine dental care such as cleanings and fillings (59). These findings, combined with the nonsignificant association between a past-year dental visit and low OHQoL, warrant further research on the costs and comprehensiveness of dental coverage for these patients.

Structural impediments were among the most common nonfinancial barriers reported by participants, including problems obtaining an appointment, not having transportation or someone to accompany them, and inconvenient clinic hours. Moreover, 28% of participants reported the lack of perceived need as a reason for not seeking care. Structural barriers may be partly mediated by oral health beliefs (60), and emphasizing the value of oral health for overall health and well-being through oral health promotion initiatives (61) may help individuals with serious mental illness overcome some of these hurdles. Nevertheless, structural barriers may prove particularly challenging. That issues beyond the control of the individual increase the likelihood of having an unmet need for dental care is alarming.

We also found that xerostomia and a schizophrenia spectrum diagnosis are independently associated with low OHQoL. Unexpectedly, the mean xerostomia score for the sample was consistent with mean scores found in nonpsychiatric samples (44). Many psychiatric medications for schizophrenia spectrum disorders have anticholinergic effects that reduce salivary flow, promoting caries and periodontal disease (62). Managing xerostomia may be challenging if high-dose medication use is indispensable, but its impact may be reduced by switching to medications with less severe anticholinergic side effects and by minimizing the risk of dental caries with regular dental visits and oral health assessments, use of sugar-free gum to stimulate salivary flow, and regular consumption of water (62). Age also appears to increase absolute risk of low OHQoL, although the association was nonsignficant in relative terms. Prior studies suggest an inconsistent relationship between age and low OHQoL but controlled for different factors (63–65); future studies can help elucidate these complex associations.

These findings suggest areas for further inquiry and possible intervention. First, integrating oral health care in their mental health care may be prudent for patients engaged in mental health services. Qualitative research suggests that individuals with serious mental illness are comfortable speaking with their mental health care providers about oral health concerns and appreciate help with dental care strategies (66). Mental health providers can more actively inquire about their patients’ oral health and medication-associated xerostomia and recommend and follow-up referrals to oral health care services when necessary. Increasingly, dental providers acknowledge the need for greater training in working with patients with mental illness (67, 68) and partnering with the mental health delivery system to integrate oral health and promote prevention within comprehensive care (69, 70).

Second, thinking dental problems were not serious or would go away on their own was a common nonfinancial barrier. Community-based health promotion may increase patients’ awareness of dental services and encourage preventive dental care (71). Third, costs were among the most frequently cited barriers to seeking care and were strongly associated with an unmet dental need in this publicly insured population (58, 59). Policy makers can decrease unmet dental need by implementing reforms in public insurance to reduce dental costs to patients and expanding dental benefits to include preventive care and more covered services (72).

Our results should be interpreted by considering several limitations. First, our exploratory study was based on a convenience sample at a public mental health service and may not be generalizable to all patients with serious mental illness. However, our sample was demographically comparable to the outpatient service population at the study site except for rates of Medicaid coverage (Medicaid only, 58% in our sample versus 38% for the study site population; Medicare-Medicaid, 28% for our sample versus 38% for the study site). Second, our results were based on self-reported data and lacked independent confirmation or comparison with a clinical assessment. Third, factors that were unexamined because of modest sample size (e.g., race-ethnicity, type of insurance) or unassessed (e.g., childhood dental care, type of dental service use) in our study merit examination in larger samples. Fourth, the cross-sectional design limits causal inferences and may contribute to the nonsignificant association between past-year dental visit and unmet dental need, given that we may have interviewed patients at varying stages of dental treatment. Fifth, although the OHIP-14 is widely used to assess OHQoL, it may underestimate the extent of low OHQoL (73), suggesting that our findings may be conservative.

Conclusions

Our study of underserved individuals receiving care for serious mental illness in a public mental health service found markedly low OHQoL, driven by unmet dental care need, xerostomia, and having a schizophrenia spectrum disorder. Financial and nonfinancial barriers were significantly associated with unmet dental need. Our findings underscore the need for further understanding and improving oral health access, oral health promotion, and greater integration between the mental and oral health systems to improve oral health in this population.