Abstract

Objective

Sleep disturbances frequently affect older adults and are linked to geriatric issues such as frailty, cognitive decline, and depression. This study assessed the psychometric properties and diagnostic value of the Turkish version of the Jenkins Sleep Evaluation Scale (JSS-TR) in individuals aged 65 years and above.

Materials and Methods

One hundred thirty-two community-dwelling older adults completed the JSS-TR along with the Pittsburgh Sleep Quality Index (PSQI), the Basic Scale on Insomnia Complaints and Quality of Sleep, and a comprehensive geriatric assessment. Analyses included internal consistency, test-retest reliability, correlation with related measures, group comparisons, and receiver operating characteristic (ROC) curve analysis to identify the most appropriate cut-off score for poor sleep quality.

Results

The median participant age was 80 years (interquartile range: 9.75). The JSS-TR demonstrated good internal consistency (Cronbach’s alpha: 0.68) and excellent stability over time (intraclass correlation coefficients: 0.92). It showed a moderate correlation with PSQI scores (r = 0.62, p < 0.001). ROC analysis indicated an optimal cutoff of ≥15, yielding 54% sensitivity and 87% specificity. Higher JSS-TR scores were significantly associated with frailty and history of falls (p < 0.05), supporting predictive validity.

Conclusion

The JSS-TR is a reliable and valid instrument for detecting poor sleep quality among older adults. A population-specific cut-off score of ≥15 improves diagnostic precision in this group.

Introduction

Sleep is a basic physiological need essential for maintaining physical and mental health. However, sleep quality tends to decline with aging. Nearly one in five older individuals is estimated to have insomnia disorder, often linked to gender, mental health, or somatic conditions (1). Worldwide, the most prevalent sleep–related disorder is obstructive sleep apnea, followed by inadequate sleep quality, other sleep problems, insomnia, and excessive daytime sleepiness (2, 3). Among community–dwelling older adults, poor sleep quality affects roughly 40% of individuals, and this prevalence may rise to 65% in institutional care (2, 3).

Sleep disturbances in older adults have multifactorial causes, including medical conditions like benign prostatic hyperplasia (through nocturia) and psychological factors such as depression and anxiety (4, 5). Sleep disturbance is closely related to geriatric syndromes as well as quality of life. For instance, poor sleep is a well–established predictor of late–life depression, conferring a 66% greater risk (4). Sleep duration shows a U–shaped relationship with frailty and depression, with both short (<6 hours) and long (>8 hours) durations, long sleep duration, in particular, has been linked to a 29% increased risk of dementia (5-8). Additionally, measures such as sleep latency, continuity, and architecture have demonstrated significant associations with cognitive performance (8). Disrupted sleep architecture is observed in individuals with mild cognitive impairment, and excessive daytime sleepiness increases the risk of cognitive decline by 26% (9, 10).

Although sleep can be measured objectively with polysomnography, it is time–consuming and impractical in most outpatient settings (6). For practical reasons, subjective tools are frequently used. Instruments such as the Pittsburgh Sleep Quality Index (PSQI) and sleep diaries are well known but may be too lengthy or complex for rapid clinical screening in geriatric practice. The Jenkins Sleep Scale (JSS) is a concise four–item questionnaire using a six–point Likert scale to assess difficulty falling asleep, waking during the night, early morning awakening, and morning tiredness (11). It has been validated in specific populations, such as cardiac and rheumatologic patients (12).

The present study aimed to assess the validity, reliability, and predictive ability of the Turkish version of the JSS in adults aged 65 years and older. Establishing its psychometric properties will support its use as a rapid, reliable, and clinically meaningful tool in geriatric practice and research.

Materials and Methods

Study Design and Participants

This cross–sectional study was conducted in geriatric outpatient clinics of two tertiary care hospitals between March and June 2025. Exclusion criteria were absence of written informed consent; communication barriers (e.g., language limitations, significant hearing loss); inability to cooperate with assessment; and neurologic or psychiatric disorders likely to affect evaluation (e.g., cerebrovascular disease, dementia, major depression, delirium). In addition, we excluded patients with acute conditions liable to cause transient sleep disturbance or whose sleep was primarily disrupted by pain or breathing difficulty at the time of assessment (e.g., acute pain exacerbation, febrile illness/infection, decompensated cardiopulmonary disease with dyspnea/orthopnea). Stable chronic comorbidities documented within the Comprehensive Geriatric Assessment (CGA) (e.g., hypertension, diabetes, osteoarthritis/chronic pain, benign prostatic hyperplasia/urinary symptoms, chronic lung disease, coronary artery disease) were permitted. Sociodemographic characteristics (age, sex, marital status, and education) were obtained via interviewer–administered questionnaires.

Sample Size

Sample size estimation followed the participant–to–item ratio described by Hogarty et al. (13), using a 20:1 ratio. For the 4–item JSS, a minimum of 80 participants was required. To account for potential data loss, 10 additional participants were recruited, resulting in a planned sample size of 90.

Instruments

Jenkins Sleep Scale (JSS)

The JSS consists of four items evaluating sleep difficulties over the preceding four weeks. Each is rated on a 6–point scale (0 = never to 5 = 22–28 days), producing a total score from 0 to 20; higher scores indicate greater disturbance (11). Scores of ≥11 suggest frequent problems. The Turkish version was previously translated and validated by Duruöz et al. (12) and permission for use was obtained.

Pittsburgh Sleep Quality Index (PSQI)

The PSQI is a 19–item measure covering seven components of sleep quality, with total scores ranging from 0 to 21. A cut–off score of ≥5 is indicative of poor sleep quality. We employed the Turkish version validated by Ağargün et al. (14).

Basic Scale on Insomnia Complaints and Quality of Sleep (BaSIQS)

BaSIQS contains seven items assessing aspects such as sleep onset latency, fragmentation, and perceived quality. Each item is rated from 0 to 4, yielding total scores of 0–28. Values above 15 indicate frequent sleep disturbances. The Turkish version validated by Ağar et al. (15) was used in the present study.

Comprehensive Geriatric Assessment (CGA)

Frailty status was determined by the Clinical Frailty Scale (CFS), with scores ≥5 defining frailty (16). Disability was defined as activities of daily living (ADL) <5 or instrumental activities of daily living (IADL) <8 (17, 18). Sarcopenia risk was screened using the SARC–F questionnaire, with scores ≥4 indicating increased risk (19). Cognitive performance was measured with the Standardized Mini–Mental State Examination, using a cut–off of <24 for impairment (20). Additional CGA components included assessments for urinary incontinence, falls within the past year, visual or hearing impairments, and nutritional status via the Mini Nutritional Assessment, with scores ≤7 indicating malnutrition (21).

Assessment of Anxiety, Depression and Stress

Anxiety symptoms were measured with the Generalized Anxiety Disorder–2 (GAD–2) scale, using a threshold of ≥3 (22). Perceived stress was assessed with the four–item Perceived Stress Scale (PSS–4), where higher scores indicate greater stress (23). Depressive symptoms were screened using the Patient Health Questionnaire–2 (PHQ–2), with a cut–off of ≥3 (24).

Cross-Cultural Adaptation

The JSS–TR, previously translated into Turkish by Duruöz et al. (12). The previously validated article employed the translation–back translation method. It was reviewed by three geriatricians to confirm its suitability for older adults. No further adaptation was deemed necessary.

Reliability Assessment

Internal consistency was examined using Cronbach’s alpha, with values above 0.6 considered acceptable. Test–retest reliability was evaluated using a two–week interval in 20 participants. Agreement for categorical items was quantified using Cohen’s kappa, interpreted as: >0.90 (almost perfect), 0.80–0.90 (strong), 0.60–0.79 (moderate), 0.40–0.59 (weak), 0.21–0.39 (minimal), and <0.20 (none). For continuous scores, intraclass correlation coefficients (ICC) were calculated using a two–way mixed–effects model (absolute agreement), with >0.90 considered excellent, 0.75–0.90 good, 0.50–0.75 moderate, and <0.50 poor.

Validity Assessment

Content validity was explored through cognitive interviews with 10 older participants to confirm clarity and relevance of items.

Construct validity was tested by examining correlations between the JSS–TR and BaSIQS (convergent validity), and between the JSS–TR and GAD–2, PHQ–2, and PSS–4 (discriminant validity).

Concurrent validity was assessed by correlation between JSS–TR and PSQI scores. Predictive validity was evaluated by comparing JSS–TR scores across subgroups with and without selected geriatric syndromes, including frailty, falls, disability, cognitive impairment, malnutrition, hearing or vision problems, and sarcopenia risk.

Agreement analysis between the JSS–TR and PSQI classifications was performed using Cohen’s kappa.

Diagnostic Accuracy

The PSQI served as the reference standard, with ≥5 indicating poor sleep quality (12). Receiver operating characteristic (ROC) curves were generated, and area under the curve (AUC) values ≥0.7 were regarded as acceptable sensitivity (Se), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV), likelihood ratios (LR⁺/LR^-), and overall accuracy were calculated with 95% confidence intervals (CIs). Three scoring approaches were examined: continuous JSS–TR score, the existing ≥11 cut–off, and the optimal cut–off (>15) determined by the Youden index. Diagnostic indices (Se, Sp, PPV, NPV, LR⁺/LR^-, AUC) were reported across plausible thresholds; results are summarized at ≥11 (screening–oriented) and ≥15 (confirmation–oriented) to reflect different clinical use cases.

Ethics approval was granted by the Clinical Research Ethics Committee of University of Health Sciences Türkiye, İzmir Tepecik Training and Research Hospital (approval number: 2025/01–14, date: 05.02.2025).

All participants provided written informed consent, and study procedures complied with the Declaration of Helsinki.

Statistical Analysis

Data analysis was performed using SPSS version 25. Normality was assessed by the Shapiro–Wilk test. Variables with normal distribution were expressed as mean ± standard deviation; non–normally distributed data were expressed as median (interquartile range [IQR]). Categorical variables were summarized as frequencies and percentages, and comparisons were made using chi–square or Fisher’s exact tests. Group differences for continuous variables were examined using independent samples t–tests or Mann–Whitney U tests, depending on distribution. Spearman’s correlation coefficients assessed associations between JSS–TR scores and other measures, with rho values interpreted. Kappa statistics were used to interpret agreement categories. ROC curve results were interpreted. The Youden index was calculated to identify optimal cut–off values. Permission to use all scales was obtained from the original authors.

Results

Participant Characteristics

A total of 132 older adults took part in the study. The median age was 80 years (IQR: 9.75), and 64.4% (n = 85) were women. Of the participants, 12.9% (n = 17) were illiterate, and 17.4% (n = 23) lived alone. The median score of CFS, ADL and IADL were 4, 6, and 8, respectively. The median score of JSS–TR, PSQI and BaSIQS were 12, 6, and 12, respectively. Sleep disturbance was identified in 52.3% (n = 69) according to the JSS–TR and in 60.6% (n = 80) based on the PSQI, as shown in Table 1.

Reliability

The JSS–TR demonstrated strong stability over time, with a kappa value of 0.806 (p < 0.001) for dichotomized scores and an ICC of 0.906 (p < 0.001) for the total score. Internal consistency reliability was acceptable for a brief 4–item multidomain screener, with a Cronbach’s alpha of 0.683.

Validity

Content Validity

Feedback from cognitive interviews with 10 older participants indicated that items were clear and relevant, with no changes required.

Construct Validity

JSS–TR scores correlated strongly with BaSIQS scores (rs = 0.624, p < 0.001). Weak but significant correlations were observed with GAD–2 (rs = 0.273, p = 0.002), PSS–4 (rs = 0.188, p = 0.03), and PHQ–2 (rs = 0.172, p = 0.05).

Predictive Validity

Higher JSS–TR scores were recorded in participants with frailty (p = 0.012), history of falls (p = 0.026), and hearing impairment (p = 0.004), as shown in Table 2. No significant associations were found for disability, sarcopenia risk, malnutrition, vision impairment, urinary incontinence, or cognitive impairment.

Concurrent Validity

JSS–TR and PSQI scores showed a moderate correlation (rs = 0.425, p < 0.001) and fair to moderate agreement (Cohen’s kappa = 0.342, p < 0.001). Agreement with BaSIQS was moderate (kappa = 0.407, p < 0.001). Table 3 presents detailed correlation results between JSS–TR and PSQI components.

Group Comparisons and Discrimination

When participants were divided into poor (n = 80) and good (n = 52) sleepers using a PSQI cut–off ≥5, total and most item scores of the JSS–TR were significantly higher in the poor sleep group (p < 0.001 for total, item 1, and item 4) (Table 4).

Diagnostic Accuracy

The ROC analysis indicated that the JSS–TR total score had an AUC of 0.706 (95% CI: 0.617–0.794) for identifying poor sleep quality (PSQI ≥5). The >15 cut–off (the confirmation–oriented threshold) demonstrated superior specificity (86.54%) and PPV (86.0%) compared to the previously recommended >11 cut–off (relative to the screening–oriented threshold), though sensitivity was lower (53.75% vs. 66.25%) [AUC = 0.701 (95% CI: 0.612–0.791), AUC = 0.677 (95% CI: 0.583–0.772), respectively]. Choice of threshold should reflect clinical intent (screening vs. confirmation), noting that NPV was modest in this cohort. Table 5 presents the full diagnostic accuracy indices, and Figure 1 shows the ROC curves for both cut–offs.

Discussion

The JSS–TR demonstrated moderate sensitivity and specificity, with a relatively high PPV, supporting its utility as a screening tool for identifying sleep disturbances among older adults. Importantly, our findings highlight the value of using a population–specific cutoff of >15, which provided greater specificity and PPV than the previously recommended cutoff of >11. The population–specific cut–off yielded a more precise identification of individuals with objectively poor sleep, as defined by the PSQI. This enhanced diagnostic accuracy is especially meaningful in geriatric care settings, where avoiding unnecessary interventions and focusing on those truly in need is a clinical priority. Our findings support the integration of the JSS–TR with the >15 cut–off as part of routine sleep assessments in older populations.

The predictive validity analysis highlighted that poor sleep, as measured by JSS–TR, was associated with frailty, falls, and hearing problems. The absence of an association with sarcopenia risk aligns with evidence indicating that while both poor sleep and reduced muscle strength are important geriatric issues, their link is likely mediated through broader health decline (6, 25). Meta–analytic data indicate that insomnia nearly doubles the risk of frailty, particularly when sleep onset and maintenance are impaired (26).

Convergent validity was supported by significant correlations with PSQI subdomains, except for “use of sleep medication,” which may be influenced by prescribing habits rather than actual sleep quality. Items related to difficulty initiating sleep and morning tiredness appeared to be the most discriminating.

Psychosocial influences on sleep in later life are well documented, with studies showing that cumulative stressors contribute to worsening sleep patterns over time (27). Our weak correlations with depression and anxiety measures suggest that JSS–TR primarily reflects sleep disturbance rather than emotional distress; however, longitudinal evidence shows a bidirectional relationship between poor sleep and depressive symptoms, underlining the importance of monitoring both in clinical care (28). Network analyses also suggest that certain JSS subdomains, such as “feeling tired in the morning,” may act as key connectors between insomnia and depression (29).

No significant association emerged between JSS–TR scores and cognitive impairment in this sample, although other research points to complex and sometimes subtle links between sleep and cognitive performance. Recent meta–analysis showed that short sleep latency and better sleep efficiency correlate with stronger executive function, while non–restorative sleep and early morning awakenings may impair cognitive flexibility and working memory (8). Differences in findings may reflect the use of self–report measures and cross–sectional design in our study.

Finally, given the simplicity, brevity, and psychometric strength of the JSS–TR, it appears suitable for routine use in geriatric outpatient settings. As sleep quality is a vital determinant of physical and psychological health in older adults (30), incorporating the JSS–TR into clinical workflows may enhance the detection and management of sleep–related problems.

Given its brevity and multidomain content, the JSS–TR is most suitable for triage and case–finding. For initial screening/triage (prioritizing sensitivity), ≥11 may be preferred. For case–finding/confirmation in geriatric clinics (prioritizing specificity and PPV), ≥15 is preferable (specificity ~87%, PPV ~86) with lower sensitivity (~54). Because NPV is modest in our sample, consistent with relatively high prevalence of poor sleep, the JSS–TR is not optimal for ruling out poor sleep on its own; positive results should be followed by confirmatory assessment (e.g., PSQI/clinical evaluation).

Study Limitations

This study has some constraints. First, its cross–sectional design limits our ability to examine changes in sleep patterns over time or to determine the scale’s sensitivity to treatment effects. Second, although we excluded individuals with diagnosed psychiatric or neurological conditions, unrecognized or subthreshold symptoms may still have influenced sleep and introduced residual bias. Third, sleep disturbances were assessed through self–report measures rather than objective tools such as actigraphy or polysomnography, which may reduce measurement precision. Fourth, while the study sample was derived from two tertiary geriatric clinics, limiting generalizability to other settings, especially community–dwelling or institutionalized older adults, future multicenter or population–based studies could provide broader validation. Finally, residual confounding from unmeasured or undiagnosed sleep–specific conditions (e.g., obstructive sleep apnea) cannot be excluded.

Conclusion

Sleep disturbances in older adults may signal underlying depression or contribute directly to symptoms such as fatigue and repeated hospital visits. Therefore, the ability to assess sleep quality rapidly and practically is essential in geriatric clinical practice. The JSS–TR proved to be a valid, reliable, and efficient screening tool for adults aged 65 years and older. The population–specific cutoff of >15 improved diagnostic precision, making it more suitable for geriatric settings than the previously recommended threshold. Future research should focus on the responsiveness of the JSS–TR to therapeutic interventions, including cognitive behavioral therapy for insomnia and pharmacologic treatments, to further establish its utility in both clinical and research settings.