Discussion

The present study examined associations among OHRQoL, emotional loneliness, subjective sleep indicators, and dream recall frequency in community-dwelling older adults. Several principal findings emerged. Better OHRQoL showed its clearest association with lower emotional loneliness. Smaller bivariate associations were observed with better subjective sleep appraisal, less frequent daytime napping, and more frequent dream recall. However, sleep quality and sleep sufficiency were almost perfectly correlated and therefore cannot be regarded as empirically distinct sleep constructs in this sample. Emotional loneliness was associated with poorer subjective sleep but was unrelated to dream recall frequency. Figure 1 provides a descriptive associative summary of the main cross-sectional relationships observed in the present study.

Figure 1: Associative framework summarizing the main observed cross-sectional relationshipsGOHAI: Geriatric Oral Health Assessment Index; Better oral health-related quality of life was associated most clearly with lower emotional loneliness, while smaller associations were observed with better subjective sleep quality and more frequent dream recall. Emotional loneliness was also associated with poorer subjective sleep quality. Gender was associated with both sleep quality and dream recall frequency, with women reporting poorer sleep quality but more frequent dream recall than men. Line thickness reflects the relative magnitude of the observed associations, and the dashed line denotes the weak, nominal, and exploratory association with dream recall. The lines are nondirectional and do not represent causal, temporal, mediational, or otherwise confirmed pathways.Image credit: Created by the authors using ChatGPT (OpenAI, California, US) and edited in Microsoft Paint (Microsoft Corp., Redmond, US).

It is not intended to represent a tested directional, causal, or mediational model.

In the adjusted models, OHRQoL remained independently associated with sleep quality and showed a small nominal association with dream recall frequency. Nevertheless, the magnitude of these associations was small, and the additional variance explained was modest, particularly for dream recall. Gender was also independently associated with both outcomes: women reported poorer subjective sleep quality but more frequent dream recall than men. In the sleep quality model, the standardized association of gender was broadly comparable in magnitude to that of OHRQoL, whereas in the dream recall model it was slightly larger. These results indicate that demographic differences were at least as relevant as OHRQoL for some of the examined outcomes.

Among the primary study variables, the strongest observed association was between OHRQoL and emotional loneliness. This finding is consistent with previous research linking oral health with psychosocial functioning and subjective well-being in later life [2,3,9]. Difficulties involving speaking, chewing, smiling, appearance, or comfort may coexist with reduced interpersonal engagement, lower self-confidence, and greater emotional disconnection. Nevertheless, the cross-sectional design does not establish whether poorer OHRQoL precedes emotional loneliness, whether loneliness affects perceptions or management of oral health, or whether both reflect shared psychological, medical, or social factors. The findings therefore support a relational interpretation of OHRQoL but do not demonstrate a causal pathway.

The distinction between emotional and social loneliness is important when interpreting this finding. Emotional loneliness concerns the perceived absence of close emotional attachment and meaningful emotional connectedness, whereas social loneliness refers more broadly to limitations in social networks and social integration [28,29]. The present results specifically concern emotional loneliness and should not be generalized to social loneliness, objective social isolation, or loneliness as a broader construct.

The association between emotional loneliness and poorer subjective sleep is consistent with evidence linking loneliness with disturbed sleep and reduced psychological well-being [17,19]. Emotional loneliness may coexist with psychological distress, emotional arousal, reduced feelings of security, and difficulty experiencing restorative sleep [17-20]. From a biopsychosocial perspective, subjective sleep quality may reflect not only physiological processes but also the emotional and relational context in which individuals live [12,13,25]. In the present study, emotional loneliness remained independently associated with sleep quality after accounting for age, gender, and OHRQoL. However, the effect was modest, and no causal or temporal conclusion can be drawn.

Disturbed sleep in later life is multifactorial and may be associated with psychological distress, medical conditions, pain, medication use, primary sleep disorders, environmental influences, and other factors not assessed in the present study. OHRQoL should therefore be regarded as one possible correlate among many rather than as a general explanation for sleep disturbance. In addition, the near-perfect correlation between sleep quality and sleep sufficiency indicates that participants did not meaningfully differentiate between these two single-item indicators. Their similar associations should therefore not be interpreted as replication across two distinct dimensions of sleep.

Gender was independently associated with both subjective sleep quality and dream recall frequency. Women reported poorer sleep quality but more frequent dream recall than men. These differences may reflect a combination of biological, psychological, social, behavioral, and reporting-related factors. However, the study was not designed to examine gender-related mechanisms, and potentially relevant explanatory variables were not assessed. The gender findings should therefore be regarded as descriptive and hypothesis-generating rather than as evidence supporting a particular explanation.

The association between OHRQoL and dream recall frequency requires substantial caution. Although the coefficient met the conventional unadjusted significance threshold after adjustment for age, gender, and emotional loneliness (β=0.091, p=0.032), the final model explained only 2.1% of the variance in dream recall frequency. The addition of OHRQoL and emotional loneliness at the second step jointly increased the explained variance by only 0.9%; this value should not be interpreted as the unique contribution of OHRQoL alone. These values indicate a very small effect with uncertain practical relevance. Given the limited explanatory power, the single nonvalidated measure of dream recall, the absence of correction for multiple comparisons, and the possibility of residual confounding, the finding is likely to be sensitive to minor changes in sample composition or model specification and may not replicate. It should therefore be considered a fragile, nominal, and hypothesis-generating observation rather than evidence of a reliable theoretical or clinically meaningful relationship.

Dream recall was assessed using a single study-specific item without established psychometric reliability or construct validity. The study did not assess dream content, vividness, emotional tone, distress, symbolism, or the broader subjective experience of dreaming. Given the limited measurement approach and the very small amount of explained variance, the observed association cannot be distinguished confidently from measurement error, individual differences in memory or reporting, residual confounding, or chance. Measurement error alone could plausibly account for the finding. It should therefore not be interpreted as evidence of a reliable, clinically meaningful, or theoretically established relationship. The result is preliminary and hypothesis-generating and requires independent replication using validated dream measures, prespecified analyses, appropriate control for multiple testing, and more comprehensive adjustment for psychological, cognitive, medical, and pharmacological factors.

Taken together, the findings are compatible with a multidimensional understanding of OHRQoL within healthy aging. Previous studies have identified associations between oral health, psychological well-being, self-care behaviors, resilience, and psychopathological symptoms [10,11]. The present findings extend this literature by examining OHRQoL alongside emotional loneliness and subjective sleep-related indicators. They should not, however, be interpreted as demonstrating that oral health determines emotional well-being, sleep quality, or dream recall. Rather, these domains may coexist within a broader biopsychosocial context in which physical, psychological, cognitive, and relational experiences interact [25].

Potential biological pathways may involve oral biofilm, periodontal inflammation, microbial processes, and associated systemic inflammatory burden [2,3]. However, the present study assessed perceived OHRQoL rather than clinical periodontal status, dental biofilm, oral infection, or inflammatory biomarkers. It therefore provides no direct evidence that biological or inflammatory mechanisms contributed to the observed associations.

From a clinical and public health perspective, OHRQoL may be considered alongside emotional and sleep-related concerns when assessing the broader well-being of older adults. The findings do not demonstrate that oral health promotion or dental treatment would reduce loneliness or improve sleep. Longitudinal and intervention studies would be required before such benefits could be inferred. Oral healthcare remains important because of its established functional, nutritional, communicative, and quality-of-life relevance, irrespective of whether it directly alters psychological or sleep-related outcomes.

Overall, the clearest finding was the association between OHRQoL and emotional loneliness. Associations involving subjective sleep were smaller, and sleep quality and sleep sufficiency did not demonstrate empirical distinctiveness. The association with dream recall frequency was weak and exploratory, while gender was at least as relevant as OHRQoL in some of the adjusted models. Given the cross-sectional design, modest effect sizes, reliance on self-report measures, potential residual confounding, and multiple statistical tests, the weaker findings require cautious interpretation and independent replication.

In later life, oral health extends beyond chewing, speaking, and freedom from pain. It is also embedded in communication, self-presentation, social participation, and subjective quality of life. The present findings indicate that oral health-related experiences were associated most clearly with emotional connectedness, whereas their smaller relationships with subjective sleep quality and dream recall remain uncertain and require further study.

Implications for person-centered care and preventive practice

The present findings have implications for person-centered care and preventive health practice, although these implications should be considered in light of the modest effect sizes and the cross-sectional design. Traditionally, oral health interventions have focused primarily on the prevention and treatment of dental disease. OHRQoL, however, also encompasses the perceived functional, psychological, and social consequences of oral conditions [2,3]. Its assessment may therefore contribute to a broader understanding of how older adults experience their health and everyday lives.

From a person-centered perspective, discussion of oral health may provide an opportunity to explore difficulties related to communication, confidence, social participation, emotional well-being, and quality of life. Clinicians should not interpret poor OHRQoL as evidence that an older person is necessarily lonely or experiencing sleep problems. Nevertheless, when oral difficulties are accompanied by social withdrawal, emotional distress, or sleep complaints, broader assessment or appropriate referral may be warranted. Comprehensive assessment can help identify coexisting needs without assuming that one domain has caused another.

The association between OHRQoL and emotional loneliness highlights the importance of considering relational and social experiences alongside physical care. Oral health difficulties may coexist with embarrassment, communication problems, reduced confidence, concern about appearance, and limitations in social participation. Emotional loneliness or reduced social engagement may likewise coexist with fewer resources, less motivation, or less support for maintaining oral health. However, these possible explanations were not directly tested, and the direction or temporal ordering of the observed relationship cannot be determined from the present study. Oral function and social participation should therefore be regarded as potentially interrelated dimensions rather than as components of a demonstrated causal sequence.

The findings do not establish that improving oral health will improve sleep quality or alter dream recall frequency. The observed sleep associations were small, while the dream recall association was particularly weak and based on a single-item measure. Consequently, dream recall should not currently be used as a clinical indicator of oral health, psychological functioning, or treatment response. Its inclusion in the present study is best understood as exploratory and as identifying a question for further research rather than an immediate target for clinical practice.

At a public health level, interdisciplinary collaboration among dental, primary care, mental health, and gerontological services may support more comprehensive care for older adults. Such collaboration is justified by the multifaceted needs of aging populations and by the established importance of oral health for daily functioning and quality of life, rather than by an assumption that oral health interventions will directly resolve loneliness or sleep difficulties. Health-promotion initiatives can appropriately situate oral health within broader strategies supporting functional ability, participation, emotional well-being, and meaningful aging [2,3].

Ultimately, person-centered care requires attention not only to disease and functional status but also to how individuals experience their bodies, relationships, and everyday lives. OHRQoL forms part of this wider experience and may help clinicians recognize concerns that would otherwise remain unspoken. However, its associations with emotional loneliness and sleep-related outcomes should be interpreted as contextual information rather than as evidence of causation or as a basis for specific psychological or sleep interventions [3-5].

Strengths of the study

The present study has several strengths, including a large community-based sample of older adults, the simultaneous examination of oral health, emotional loneliness, sleep experiences, and dream recall within a single framework, and the use of validated measures of OHRQoL and loneliness.

Study limitations and future directions

Several limitations should be considered when interpreting the findings. First, the cross-sectional design does not permit conclusions regarding causality, temporal ordering, or directionality. Although OHRQoL was statistically associated with emotional loneliness, sleep quality, and dream recall frequency, the findings do not demonstrate that poorer OHRQoL causes loneliness or sleep-related difficulties. Reverse or reciprocal relationships are also possible, and the observed associations may reflect shared underlying factors.

Second, the associations with the sleep-related variables were generally small. The association between OHRQoL and emotional loneliness was the strongest of the primary associations, whereas those with sleep quality, daytime napping, and dream recall frequency were considerably weaker. The regression models also explained only a modest proportion of outcome variance. Statistical significance in a relatively large sample should not be interpreted as evidence of clinical or practical importance. The practical relevance of the sleep-related findings therefore remains uncertain.

Third, the original sample-size determination was based on a precision-based estimate for a population proportion rather than on a prespecified effect size for the correlational and regression analyses reported in this study. Accordingly, although the final analytic sample exceeded the recruitment target, the calculation should not be interpreted as a formal a priori power analysis for the specific statistical models examined.

Fourth, the association with dream recall frequency was particularly small and was based on a single study-specific item with no established psychometric reliability or construct validity. The final regression model explained only 2.1% of the variance in dream recall frequency, while the variables introduced at the second step jointly increased the explained variance by only 0.9%. These values indicate minimal explanatory value. Given the limited measurement approach and the very small amount of explained variance, the observed association cannot be distinguished confidently from measurement error, individual differences in reporting or memory, residual confounding, sampling variation, or chance. Measurement error alone could plausibly account for the finding. The OHRQoL coefficient may also be unstable and could change or become nonsignificant with minor variations in sample composition, covariate selection, measurement, or model specification. Because no formal robustness or specification-sensitivity analyses were conducted, the stability of this finding cannot be established. In addition, the study did not assess dream content, vividness, emotional tone, distress, symbolism, or other dimensions of dreaming. Consequently, the finding should not be interpreted as evidence concerning the broader experience or psychological function of dreaming. Its theoretical and clinical meaning remains uncertain, and it should be treated solely as preliminary and hypothesis-generating.

Fifth, although age and gender were included as covariates and emotional loneliness was considered in the primary regression models, several potentially important confounding variables were not included in the analyses. These include depressive and anxiety symptoms, medication use, pain, diagnosed sleep disorders, multimorbidity, socioeconomic circumstances, and recent dental treatment. Clinical periodontal status, dental biofilm, specific oral and dental conditions, markers of oral infection, and systemic inflammatory biomarkers were also not assessed. Residual confounding may therefore have contributed to some of the observed associations, and the study cannot determine whether biological or inflammatory processes were involved.

Cognitive functioning was not assessed using a standardized screening instrument. Consequently, milder or unrecognized cognitive impairment may have affected participants’ comprehension, recall, and response reliability, particularly for the single-item measure of dream recall frequency. At the same time, the requirement that participants complete the questionnaire independently may have resulted in the underrepresentation of older adults with more substantial cognitive impairment.

Sixth, several bivariate associations and regression models were examined. Because no formal adjustment for multiple comparisons was applied, the possibility of Type I error and false-positive findings cannot be excluded, particularly for the smaller associations. The weaker findings, especially those involving dream recall frequency, should therefore be interpreted as exploratory and require confirmation in independently recruited samples. Moreover, because the theory-guided expectations and analytical distinctions were not preregistered, the findings should not be regarded as strictly confirmatory.

Finally, participants were community-dwelling older adults recruited in Greece through community centers, local health units, social services, and community associations using a non-probability convenience sampling strategy. Individuals who attend such settings may differ systematically from the wider older adult population. In particular, they may be more socially connected, functionally independent, or engaged with community and healthcare services, although those recruited through health units may also have had greater health needs. The sample may therefore have underrepresented homebound, highly isolated, frail, institutionalized, or cognitively impaired older adults. This selection process may have affected the observed distribution of emotional loneliness and restricted variability in social and health-related characteristics, potentially influencing the magnitude of the reported associations. The direction of this bias cannot be determined from the available data. Consequently, the findings should not be generalized to the broader older adult population or to older adults living in different cultural and healthcare contexts.

Future research should employ longitudinal or prospective designs, validated multidimensional measures of sleep and dreaming, more detailed assessments of daytime napping, and clinical oral and periodontal examinations. More comprehensive adjustment for psychological, medical, pharmacological, cognitive, and socioeconomic factors would also be needed. The inclusion of microbiological or inflammatory biomarkers could further support the investigation of possible biological mechanisms underlying the observed associations. Preregistered hypotheses, prespecified analytical plans, and appropriate correction for multiple testing would strengthen the reliability and interpretability of future findings.

Future studies could additionally examine the emotional tone, imagery, recurring themes, and personally attributed meaning of dreams. Such research might explore whether distressing dreams are experienced as subjective signals of psychological strain or of waking-life concerns requiring attention, including, but not limited to, health and self-care needs. This possibility could not be evaluated in the present study, which assessed only dream recall frequency and did not examine dream content, emotional valence, or distress.