Abstract
Sleep disturbances are highly prevalent among elderly patients with Parkinson disease (PD), substantially impairing their quality of life. This study aimed to identify associated risk factors and to develop and validate a nomogram risk prediction model based on multidimensional data, providing a practical tool for early clinical identification and intervention. A retrospective case-control study was conducted on 340 elderly PD patients admitted to a tertiary hospital between June 2023 and June 2025. Multidimensional data were collected through self-designed questionnaires, standardized clinical scales, and electronic medical records. Patients were classified into a sleep disorder group and a non-sleep disorder group using the second version of the Parkinson Disease Sleep Scale (PDSS-2) with a cutoff score of ≥18. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors, which were subsequently incorporated into the nomogram model. Model discrimination was assessed by the receiver operating characteristic (ROC) curve, while calibration was evaluated using bootstrap resampling and calibration plots. The prevalence of sleep disturbances in elderly hospitalized PD patients was 61.2%. Multivariate analysis identified advanced Hoehn–Yahr stage, higher UPDRS-II score, severe nocturnal pain, increased nocturia frequency, and elevated HADS-depression scores as independent predictors of sleep disturbances. The developed nomogram demonstrated good discriminative ability, with an area under the ROC curve (AUC) of 0.867, sensitivity of 0.884, and specificity of 0.765. The Hosmer–Lemeshow test indicated satisfactory model calibration. Sleep disturbances are common in elderly hospitalized PD patients and are influenced by multiple clinical factors. The proposed nomogram model exhibits favorable predictive performance and calibration, offering an effective clinical tool for risk stratification and facilitating early intervention.
Keywords: multidimensional data, nomogram, Parkinson disease, risk prediction model, sleep disorders
1. Introduction
Parkinson disease (PD) is the second most prevalent neurodegenerative disorder after Alzheimer disease, affecting more than 6 million individuals worldwide.[1] PD prevalence rises sharply with age, reaching approximately 1% to 2% in adults over 65 years and 3% to 4% in those over 85 years.[2] Driven by global population aging, the total number of PD patients is projected to exceed 12 million by 2040, with older adults accounting for the majority of cases, posing substantial challenges to geriatric healthcare systems.[3] Beyond classic motor symptoms (bradykinesia, tremor, rigidity, and postural instability), elderly PD patients frequently experience age-related non-motor symptoms, among which sleep disturbances are particularly common and clinically significant.[4] The prevalence of sleep disorders in this population exceeds 60%,[5] markedly higher than in age-matched non-PD controls, and stems from the combined effects of age-related sleep architecture changes and PD-related neurodegeneration of sleep–wake regulatory circuits.[6]
Sleep disturbances in elderly PD patients present with age-specific features, including insomnia, excessive daytime sleepiness, rapid eye movement sleep behavior disorder (RBD), restless legs syndrome, sleep-disordered breathing, and circadian rhythm disruption.[6] These impairments exert multifaceted adverse effects: they reduce quality of life via daytime fatigue, reduced concentration, and cognitive decline[7]; exacerbate motor symptoms such as gait and balance instability, increasing fall risk[8]; and are closely associated with anxiety and depression, which may accelerate cognitive deterioration and dementia progression.[9] Notably, as a prodromal marker of PD, RBD carries additional physical injury risk for elderly patients with osteoporosis due to vigorous dream-enactment behaviors.[8]
Despite their high prevalence and clinical burden, sleep disturbances in elderly PD patients remain underrecognized in clinical practice. Comorbidities and polypharmacy complicate routine assessment. Subjective questionnaires such as the Pittsburgh Sleep Quality Index are widely used due to their simplicity, but their accuracy is compromised by cognitive impairment and recall bias in older adults.[10] Polysomnography (PSG), the diagnostic gold standard, provides objective data but is limited by high cost, technical demands, and poor feasibility for frail or immobile older patients, making it unsuitable for routine screening.[11] These barriers delay early intervention, highlighting the need for a practical, objective risk assessment tool tailored to this population. Nomograms, as intuitive graphical prediction models derived from multivariable regression, enable rapid individualized risk estimation and support clinical decision-making.[12]
Accordingly, this study aims to develop and validate a nomogram risk prediction model for sleep disorders in elderly PD patients by integrating multidimensional clinical, scale-based, and electronic medical record data. The model is intended to support early identification and personalized management of sleep disturbances, ultimately improving patient quality of life and advancing evidence-based precision care for elderly PD populations.
2. Materials and methods
2.1. Study population
This retrospective case-control study included patients with Parkinson disease (PD) who received diagnosis and treatment at our hospital between June 2023 and June 2025. The inclusion criteria were as follows: age ≥ 60 years; diagnosis consistent with the clinical diagnostic criteria of the International Parkinson and Movement Disorder Society[13]; complete medical records; and good treatment compliance. Exclusion criteria included: history of cognitive impairment or psychiatric disorders; coexisting malignant tumors; and severe dysfunction of vital organs such as the heart, lungs, brain, or kidneys.
The number of variables included in the study was 17. According to the principle that sample size should be 5 to 10 times the number of variables, and based on the reported prevalence of sleep disorders in PD patients (63.7%),[14] the minimum sample size was calculated as 334 cases after accounting for a 20% data loss rate (17 × 10/0.637/0.8). The final sample size was determined to be 340 cases.
This study was conducted in accordance with the Declaration of Helsinki. The protocol was reviewed and approved by the Ethics Committee of The Affiliated Brain Hospital of Nanjing Medical University (approval number: 2025-1224). Due to the retrospective nature and use of anonymized data, the requirement for informed consent was waived.
2.2. Research instruments
2.2.1. Parkinson Disease Sleep Scale-2 (PDSS-2)
The PDSS-2 is an internationally recognized instrument for assessing sleep disturbances in PD. It consists of 15 items across 3 domains, each rated on a 5-point Likert scale (0–4), yielding a total score range of 0 to 60, with higher scores indicating more severe sleep problems. The PDSS-2 has been validated in multiple countries, with Cronbach α ranging from 0.81 to 0.94.[15] A cutoff score of ≥18 points was applied to define clinically significant sleep disturbances. Patients were therefore categorized into the sleep disorder group (PDSS-2 ≥ 18) and the non-sleep disorder group (PDSS-2 < 18).
2.2.2. Hospital Anxiety and Depression Scale (HADS)
The HADS is a widely used screening tool for anxiety and depression in nonpsychiatric populations. It comprises 14 items, divided into an anxiety subscale and a depression subscale (7 items each). Each item is rated on a 4-point scale (0–3), with subscale scores ranging from 0 to 21. Higher scores reflect more severe symptoms: <7 = normal, 8 to 10 = borderline abnormal, and 11 to 21 = definite abnormal. The HADS demonstrates good reliability, with total Cronbach α = 0.85; the anxiety and depression subscales have α coefficients of 0.79 and 0.82, respectively.[16]
2.2.3. Unified Parkinson Disease Rating Scale-II (UPDRS-II)
The UPDRS-II is one of the gold-standard instruments for evaluating functional status in PD patients. It contains 13 items addressing speech, salivation, swallowing, cutting food, dressing, hygiene, and other daily activities. Each item is rated from 0 to 4, producing a total score of 0 to 52, with higher scores reflecting more severe functional impairment. The UPDRS-II has shown good internal consistency (Cronbach α = 0.87).[17]
2.2.4. Nocturnal pain assessment
Nocturnal pain was assessed using the Visual Analogue Scale (VAS).[18] Scores ranged from 0 to 10, with higher scores indicating greater pain intensity: 0 = no pain; 1 to 3 = mild pain (tolerable, with minimal sleep disturbance); 4 to 6 = moderate pain (affecting sleep quality, sometimes awakening the patient); and 7 to 10 = severe pain (intolerable, preventing sleep).
2.3. Observation indicators
A self-designed questionnaire for assessing risk factors of sleep disturbances in elderly PD patients was developed with input from neurology and geriatrics specialists and by reviewing relevant literature. The questionnaire included:
Demographics and comorbidities: age, sex, body mass index (BMI), educational level, marital status, smoking history, alcohol consumption, and number of comorbidities.
Clinical characteristics: Hoehn–Yahr stage, disease duration, UPDRS-II score, nocturnal pain (VAS), nocturia frequency, and use of sedative/hypnotic medications.
Psychological features: HADS scores.
All data were extracted from the hospital information system (HIS) by 2 trained researchers and independently cross-checked by a third investigator to ensure accuracy and completeness. As specified in the inclusion criteria, all enrolled patients had complete medical records with no missing data for the key variables; therefore, no imputation methods were applied.
2.4. Statistical analysis
Statistical analyses were performed using SPSS version 23.0 (IBM Corporation) and R version 4.4.2 (The R Foundation for Statistical Computing). Categorical variables were expressed as frequencies and percentages, and group comparisons were conducted using the χ2 test. Variables with statistical significance in univariate analysis (P < .05) were entered into a multivariable logistic regression model, and a nomogram prediction tool was developed based on regression coefficients. The discriminative performance of the model was evaluated by receiver operating characteristic (ROC) curve analysis, and calibration was assessed using the Hosmer–Lemeshow goodness-of-fit test. A 2-sided P-value < .05 was considered statistically significant.
3. Results
3.1. Comparison of baseline characteristics
A total of 340 patients with PD were included, comprising 208 patients in the sleep disorder group and 132 in the non-sleep disorder group. Univariate analysis indicated significant between-group differences in Hoehn–Yahr stage, UPDRS-II score, nocturnal pain severity, nocturia frequency, and HADS scores (P < .05). No statistically significant differences were observed in age, sex, body mass index (BMI), educational level, marital status, smoking history, alcohol consumption, number of comorbidities, disease duration, or use of hypnotic/sedative medications (P > .05) (Table 1).
Table 1.
Comparison of general patient characteristics between 2 groups, n (%).
| Clinical data | Sleep disorder group (n = 208) | Non-sleep disorder group (n = 132) | χ2/t-value | P-value |
|---|---|---|---|---|
| Age (yr) | 3.485 | .1752 | ||
| ≤65 | 90 (43.3) | 65 (49.2) | ||
| 66–75 | 75 (36.1) | 50 (37.9) | ||
| ≥76 | 43 (20.7) | 17 (12.9) | ||
| Gender | 0.288 | .591 | ||
| Male | 112 (53.8) | 75 (56.8) | ||
| Female | 96 (46.2) | 57 (43.2) | ||
| BMI (kg/m2) | 2.650 | .266 | ||
| <18.5 | 21 (10.1) | 8 (6.1) | ||
| 18.5–23.9 | 95 (45.7) | 70 (53.0) | ||
| ≥24 | 92 (44.2) | 54 (40.9) | ||
| Educational attainment | 3.975 | .137 | ||
| Elementary school or below | 78 (37.5) | 41 (31.1) | ||
| Junior high school | 85 (40.9) | 50 (37.9) | ||
| High school or above | 45 (21.6) | 41 (31.1) | ||
| Marital status | 0.002 | .961 | ||
| Married | 165 (79.3) | 105 (79.5) | ||
| Single | 43 (20.7) | 27 (20.5) | ||
| Smoking history | 0.003 | .954 | ||
| Yes | 75 (36.1) | 48 (36.4) | ||
| No | 133 (63.9) | 84 (63.6) | ||
| Alcohol consumption history | 0.124 | .725 | ||
| Yes | 62 (29.8) | 37 (28.0) | ||
| No | 146 (70.2) | 95 (72.0) | ||
| Number of comorbidities (types) | 0.968 | .616 | ||
| 0–1 | 87 (41.8) | 62 (47.0) | ||
| 2–3 | 98 (47.1) | 58 (43.9) | ||
| ≥4 | 23 (11.1) | 12 (9.1) | ||
| Hoehn–Yahr stages | 20.914 | .002 | ||
| 1–2 | 68 (32.7) | 75 (56.8) | ||
| 3 | 97 (46.6) | 45 (34.1) | ||
| 4–5 | 43 (20.7) | 12 (9.1) | ||
| Duration of illness (yr) | 5.687 | .061 | ||
| ≤3 | 75 (36.1) | 62 (47.0) | ||
| 4–7 | 87 (41.8) | 52 (39.4) | ||
| ≥8 | 46 (22.1) | 18 (13.6) | ||
| UPDRS-II score (points) | 35.082 | <.001 | ||
| ≤10 | 55 (26.4) | 75 (56.8) | ||
| 11–20 | 98 (47.1) | 45 (34.1) | ||
| ≥21 | 55 (26.4) | 12 (9.1) | ||
| Nighttime pain level | 58.307 | <.001 | ||
| No pain | 35 (16.8) | 65 (49.2) | ||
| Mild | 78 (37.5) | 52 (39.4) | ||
| Moderate to severe | 95 (45.7) | 15 (11.4) | ||
| Frequency of nocturia (times/night) | 84.469 | <.001 | ||
| 0–1 | 45 (21.6) | 95 (72.0) | ||
| ≥2 times | 163 (78.4) | 37 (28.0) | ||
| Taking hypnotics/sedatives | 2.913 | .088 | ||
| Yes | 65 (31.3) | 30 (22.7) | ||
| No | 143 (68.8) | 102 (77.3) | ||
| HADS anxiety score (points) | 6.527 | .038 | ||
| <7 | 72 (34.6) | 55 (41.7) | ||
| 8–10 | 80 (38.5) | 57 (43.2) | ||
| ≥11 | 56 (26.9) | 20 (15.2) | ||
| HADS depression score (points) | 55.616 | <.001 | ||
| <7 | 65 (31.3) | 95 (72.0) | ||
| 8–10 | 75 (36.1) | 25 (18.9) | ||
| ≥11 | 68 (32.7) | 12 (9.1) |
BMI = body mass index, HADS = Hospital Anxiety and Depression Scale, UPDRS-II = Unified Parkinson Disease Rating Scale Part II, VAS = Visual Analogue Scale.
3.2. Multivariate logistic regression analysis
Sleep disorder (defined as PDSS-2 ≥ 18) was used as the dependent variable, and variables with P < .05 in univariate analysis were included as independent variables: Hoehn–Yahr stage, UPDRS-II score, nocturnal VAS score, nocturia frequency, HADS-anxiety score, and HADS-depression score. HADS-anxiety did not reach statistical significance (P = .253). Multivariate logistic regression identified higher Hoehn–Yahr stage, higher UPDRS-II score, more severe nocturnal pain, increased nocturia frequency, and higher HADS-depression scores as independent risk factors for sleep disorder in elderly PD patients. The regression equation was: Y = 0.352 + 0.284 × Hoehn–Yahr stage + 0.613 × UPDRS-II score + 0.710 × nocturnal pain + 1.638 × nocturia frequency + 0.418 × HADS-depression score (Tables 2 and 3). Exploratory analyses of multiplicative interaction terms (e.g., HADS-depression × Hoehn–Yahr stage, HADS-depression × UPDRS-II score) were performed, and none reached statistical significance after Bonferroni correction (all P > .10).
Table 2.
Variable assignment status.
| Variables | Scoring explanation |
|---|---|
| Presence of sleep disorders | PDSS-2 < 18 points = 0, PDSS-2 ≥ 18 points = 1 |
| Hoehn–Yahr stage | Stages 1–2 = 0, Stage 3 = 1, Stages 4–5 = 2 |
| UPDRS-II score | ≤10 points = 0, 11–20 points = 1, ≥21 points = 2 |
| Nighttime pain severity | No pain = 0, mild = 1, moderate to severe = 2 |
| Nocturia frequency | 0–1 times/night = 0, ≥2 times/night = 1 |
| HADS anxiety score | <7 points = 0, 8–10 points = 1, ≥11 points = 2 |
| HADS depression score | <7 points = 0, 8–10 points = 1, ≥11 points = 2 |
HADS = Hospital Anxiety and Depression Scale, PDSS-2 = Parkinson Disease Sleep Scale-2, UPDRS-II = Unified Parkinson Disease Rating Scale Part II, VAS = Visual Analogue Scale.
Table 3.
Multivariate logistic regression analysis of sleep disorders in elderly patients with Parkinson disease.
| Variables | β | SE | Wald χ2 | P-value | OR (95% CI) |
|---|---|---|---|---|---|
| Constant | 0.352 | 0.245 | 2.064 | .151 | – |
| Hoehn–Yahr staging | 0.284 | 0.130 | 4.753 | .029 | 1.328 (1.029–1.714) |
| UPDRS-II score | 0.613 | 0.185 | 10.938 | .001 | 1.846 (1.284–2.656) |
| Nighttime pain intensity | 0.710 | 0.194 | 13.320 | <.001 | 2.033 (1.389–2.977) |
| Nocturia frequency | 1.638 | 0.332 | 24.350 | <0.001 | 5.146 (2.685–9.864) |
| HADS depression score | 0.418 | 0.189 | 4.884 | .027 | 1.520 (1.048–2.203) |
| HADS-anxiety score | 0.352 | 0.308 | 1.306 | .253 | 1.422 (0.777–2.600) |
CI = confidence interval, HADS = Hospital Anxiety and Depression Scale, OR = odds ratio, SE = standard error, UPDRS-II = Unified Parkinson Disease Rating Scale Part II, β = regression coefficient.
3.3. Construction of the nomogram prediction model
Based on the regression results, a nomogram was developed using R software to predict the risk of sleep disorder in elderly PD patients. The nomogram incorporated Hoehn–Yahr stage, UPDRS-II score, nocturnal pain, nocturia frequency, HADS-depression score, and predicted probability. To use the nomogram, 1 can draw lines from each predictor’s value to the “Points” axis, sum the points, and then project the total onto the “Predicted Probability” axis (see Fig. 1 legend for a detailed example).
Figure 1.
Risk prediction model for sleep disorders in elderly Parkinson patients. To use the nomogram, locate the patient’s value on each predictor axis (Hoehn–Yahr stage, UPDRS-II score, nocturnal pain VAS, nocturia frequency, HADS-depression score). Draw a vertical line upward to the “Points” axis to obtain the score for each variable. Sum the scores to get “Total Points.” Draw a vertical line downward from the “Total Points” axis to the “Predicted Probability” axis to read the estimated risk of sleep disorder (PDSS-2 ≥ 18). An example patient (red dots) with total points of 122 corresponds to a predicted risk of approximately 65%. HADS = Hospital Anxiety and Depression Scale, PDSS-2 = Parkinson Disease Sleep Scale-2, UPDRS-II = Unified Parkinson Disease Rating Scale Part II, VAS = Visual Analogue Scale.
3.4. Model validation
ROC curve analysis demonstrated good discriminative performance of the model, with an AUC of 0.867 (95% CI, 0.826–0.908; Fig. 2). The optimal cutoff, determined by the maximum Youden index, was 0.629, corresponding to a sensitivity of 0.884 and specificity of 0.765, with a predicted probability threshold of 0.464. Internal validation using bootstrap resampling (1000 iterations) showed good calibration, with the Hosmer–Lemeshow test indicating χ2 = 8.341, P = .207. The calibration curve closely approximated the ideal line, confirming good model fit (Fig. 3).
Figure 2.
ROC curve of the nomogram risk prediction model. ROC = receiver operating characteristic.
Figure 3.
Calibration curve of the nomogram risk prediction model. ROC = receiver operating characteristic.
4. Discussion
4.1. High Prevalence and Multifactorial Influences
In this study of 340 hospitalized elderly PD patients, the prevalence of sleep disorder was 61.2%, substantially higher than that reported in the general elderly population and in non-PD chronic disease cohorts.[19] Multivariate analysis revealed that advanced Hoehn–Yahr stage, higher UPDRS-II score, nocturnal pain, nocturia frequency, and depressive symptoms were independent risk factors. Consistent with previous studies,[20] our findings confirm the negative impact of disease severity and motor dysfunction on sleep, while also highlighting the critical roles of nocturnal pain, urinary symptoms, and depression. Notably, nocturia had the strongest effect (OR = 5.146), underscoring the importance of screening and managing urinary symptoms in clinical care. These findings emphasize the need for comprehensive, multidimensional assessment and early intervention in elderly PD patients. Although no significant sex difference was observed between the 2 groups in the present study (P = .591), accumulating evidence indicates that female PD patients are more likely to experience sleep disturbances and depressive symptoms than their male counterparts.[21] Thus, the lack of a significant sex effect might be attributable to limited statistical power rather than a true absence of sex differences. Furthermore, detailed information on the types and doses of dopaminergic medications (e.g., levodopa equivalent daily dose) was not analyzed, which may have introduced residual confounding, as dopamine replacement therapy is known to affect sleep architecture and nocturia frequency. These limitations should be considered when interpreting the findings.
4.2. Predictive value and clinical applicability of the nomogram
Compared with using the PDSS-2 alone (reported AUC 0.79, sensitivity 76.0%, specificity 69.0%),[22] our nomogram demonstrated better discriminative performance (AUC = 0.867, sensitivity 0.884, specificity 0.765), surpassing previous models based solely on clinical history or limited variables.[23] Internal validation confirmed robust calibration and stability. By integrating multiple independent predictors, the model avoided redundancy and maintained simplicity, ensuring strong clinical applicability. The visual scoring system allows healthcare providers to rapidly estimate individual risk and tailor interventions accordingly, reinforcing the role of multidimensional data in precision nursing. The nomogram is therefore recommended for routine inpatient assessments to facilitate early identification and stratified management of high-risk patients.
4.3. Pathophysiological mechanisms and multidimensional factors
4.3.1. Disease severity and motor dysfunction
Advanced Hoehn–Yahr stage (OR = 1.328) and higher UPDRS-II scores (OR = 1.846) were strongly associated with sleep disturbance.[24] Neurodegeneration in brainstem sleep–wake centers, dopaminergic neuronal loss, and motor symptoms such as rigidity and tremor contribute to fragmented sleep.[25] Functional impairment further reduces daytime activity, disrupting circadian rhythm. Comprehensive rehabilitation, including balance and gait training with optimized scheduling, may improve nocturnal sleep quality.
4.3.2. Nocturnal pain and nocturia
Nocturnal pain (OR = 2.033) and nocturia (OR = 5.146) emerged as major predictors. Pain activates the hypothalamic–pituitary–adrenal axis, increasing cortisol and proinflammatory cytokines, thereby disrupting non-REM sleep.[26] Nocturia is largely attributed to autonomic dysfunction leading to reduced bladder capacity and detrusor overactivity, compounded by dopaminergic medications.[24] Despite their prevalence, these symptoms are often underrecognized. Comprehensive management combining pharmacologic and non-pharmacologic measures (e.g., timed voiding, pain-focused CBT, stepwise analgesia, anticholinergic agents) should be incorporated into care plans.
4.3.3. Depressive symptoms
HADS-depression score (OR = 1.520) independently predicted sleep disturbance. Dysregulation of serotonergic and noradrenergic systems and frontolimbic connectivity abnormalities contribute to both depression and poor sleep.[27] Although HADS-anxiety showed a significant association in univariate analysis (P = .038), this effect did not remain significant after adjusting for depression and other clinical variables (P = .253), likely due to the high collinearity between anxiety and depression in this population (Spearman ρ = 0.62). Nevertheless, the frequent co-occurrence of anxiety and depression suggests synergistic effects on sleep outcomes. Early identification and targeted interventions – such as mindfulness-based therapy, psychosocial support, and antidepressants where appropriate – are critical.[28] Nurses should actively screen using HADS and coordinate psychological or psychiatric consultations when indicated.
4.4. Strengths and limitations
This study is the first to construct a multidimensional nomogram incorporating motor, somatic, urinary, and psychological dimensions for predicting sleep disturbance in elderly PD patients. The model demonstrated superior discrimination and calibration (AUC = 0.867), offering a practical tool for precision nursing. However, several limitations should be acknowledged. First, the single-center, retrospective design and the inclusion of only hospitalized patients may limit generalizability. Hospitalized elderly PD patients likely have a more severe disease spectrum than community-dwelling counterparts, potentially inflating the observed prevalence of sleep disorders. Second, despite incorporating multidimensional clinical data, objective measures such as neuroimaging, electrophysiological, or molecular biomarkers were not included. Neuroimaging studies have demonstrated that gray matter atrophy in frontotemporal and subcortical regions, as well as cholinergic system degeneration, is associated with non-motor symptoms in PD,[15,29] and sleep disorders have been linked to cognitive dysfunction.[30] However, the absence of such data in our model limits the ability to explore the neural mechanisms underlying sleep disorders. Third, external validation with an independent cohort was not conducted in this study. Restricted by the single-center retrospective design and strict inclusion/exclusion criteria (age ≥ 60 years, complete medical records, no history of cognitive impairment or malignant tumors, etc), no independent patient cohort meeting the same standardized criteria was available within the study period. Predictive models developed and validated within the same dataset are prone to overfitting, and their performance may deteriorate when applied to new populations. Despite these limitations, our findings provide a readily available assessment tool and actionable intervention targets for clinical nursing. Future multicenter prospective studies incorporating neuroimaging and molecular markers are warranted to enhance generalizability and predictive accuracy.
5. Conclusion
In this study, we successfully developed and validated a nomogram prediction model for sleep disorders in elderly patients with Parkinson disease based on multidimensional data. The model demonstrated excellent discriminative ability and calibration, with an AUC of 0.867, highlighting its potential as a reliable and practical tool for clinical use. By integrating key predictors – Hoehn–Yahr stage, UPDRS-II score, nocturnal pain, nocturia frequency, and depressive symptoms – the nomogram enables healthcare providers to efficiently identify high-risk individuals and implement timely, personalized interventions. This tool supports a proactive and stratified approach to managing sleep disturbances in elderly PD patients, ultimately contributing to improved quality of life and reduced disease burden. Future efforts should focus on external validation across diverse populations, particularly community-based cohorts, and the integration of additional objective biomarkers (e.g., neuroimaging, polysomnography) to further enhance predictive accuracy and generalizability in the context of precision medicine.
Acknowledgments
The authors thank all patients and their families for their participation and cooperation in this study, as well as all medical and nursing staff of the Department of Geriatrics at The Affiliated Brain Hospital of Nanjing Medical University for their support.
Author contributions
Conceptualization: Yanyan Wu.
Data curation: Linlin Wu.
Formal analysis: Linlin Wu.
Funding acquisition: Yanyan Wu.
Methodology: Linlin Wu, Li Zhang.
Project administration: Yanyan Wu.
Resources: Yanyan Wu.
Writing – original draft: Linlin Wu.
Writing – review & editing: Linlin Wu, Yanyan Wu.
Abbreviations:
- AUC
- area under the curve
- CI
- confidence interval
- HADS
- Hospital Anxiety and Depression Scale
- HIS
- hospital information system
- OR
- odds ratio
- PD
- Parkinson disease
- PDSS-2
- Parkinson Disease Sleep Scale-2
- ROC
- receiver operating characteristic
- UPDRS-II
- Unified Parkinson Disease Rating Scale Part II
- VAS
- Visual Analogue Scale
The authors have no funding and conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
How to cite this article: Wu L, Zhang L, Wu Y. Development and validation of a nomogram prediction model for sleep disorders in elderly Parkinson disease patients based on multidimensional data: A retrospective case-control study. Medicine 2026;105:27(e49580).
Contributor Information
Linlin Wu, Email: 13813832216@163.com.
Li Zhang, Email: neuro_zhangli@163.com.
References
- [1].Ou Z, Pan J, Tang S, et al. Global trends in the incidence, prevalence, and years lived with disability of Parkinson’s disease in 204 countries/territories from 1990 to 2019. Front Public Health. 2021;9:776847. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Muleiro Alvarez M, Cano-Herrera G, Osorio Martínez MF, et al. A Comprehensive approach to Parkinson’s disease: addressing its molecular, clinical, and therapeutic aspects. Int J Mol Sci. 2024;25:7183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Zhang F, Liu YX, Zhu YY, et al. Epidemiologic risk and prevention and interventions in Parkinson disease: from a nutrition-based perspective. J Nutr. 2025;155:1019–30. [DOI] [PubMed] [Google Scholar]
- [4].Xu Z, Anderson KN, Pavese N. Longitudinal studies of sleep disturbances in Parkinson’s disease. Curr Neurol Neurosci Rep. 2022;22:635–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Maggi G, Vitale C, Cerciello F, Santangelo G. Sleep and wakefulness disturbances in Parkinson’s disease: a meta-analysis on prevalence and clinical aspects of REM sleep behavior disorder, excessive daytime sleepiness and insomnia. Sleep Med Rev. 2023;68:101759. [DOI] [PubMed] [Google Scholar]
- [6].Iranzo A, Cochen De Cock V, Fantini ML, Pérez-Carbonell L, Trotti LM. Sleep and sleep disorders in people with Parkinson’s disease. Lancet Neurol. 2024;23:925–37. [DOI] [PubMed] [Google Scholar]
- [7].Germann J, Neumann S. Editorial: the neuroscience of Parkinson’s disease: exploring causes, symptoms, and potential treatments. Front Hum Neurosci. 2025;19:1609635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Prins EM, Pijpers A, Verbeek N, et al. The impact of REM-sleep behavior disorder on people with Parkinson’s disease and their partners. npj Parkinson’s Dis. 2025;11:130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Asadpoordezaki Z, Henley BM, Coogan AN. Prevalence and associations of self-reported sleep problems in a large sample of patients with Parkinson’s disease. J Sleep Res. 2025;34:e14453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Marafioti G, Corallo F, Cardile D, Lorenzo G D, Quartarone A, Buono V L. REM sleep behavior disorder and cognitive functions in Parkinson’s patients: a systematic review. J Clin Med. 2023;12:7397. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Mureșan RC, Vulturar DM, LȘ Moacă, et al. Current aspects and future perspectives on polysomnography. Applicability in rehabilitation patients diagnosed with sleep related breathing disorders – a narrative review. Balneo PRM Res J. 2024;15:659. [Google Scholar]
- [12].Bugalho P, Magriço M, Alves L, Borbinha C. Objective sleep data as predictors of cognitive decline in dementia with lewy bodies and Parkinson’s disease. Sleep Med. 2021;80:273–8. [DOI] [PubMed] [Google Scholar]
- [13].Montanaro E, Romagnolo A, Fabbri M, et al. Association between sleep disorders and cognitive dysfunctions in non-demented patients with advanced Parkinson’s disease. J Neurol. 2022;269:1538–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Dijkstra F, de Volder I, Viaene M, Cras P, Crosiers D. Polysomnographic predictors of sleep, motor, and cognitive dysfunction progression in Parkinson’s disease. Curr Neurol Neurosci Rep. 2022;22:657–74. [DOI] [PubMed] [Google Scholar]
- [15].Maggi G, Trojano L, Barone P, Santangelo G. Sleep disorders and cognitive dysfunctions in Parkinson’s disease: a meta-analytic study. Neuropsychol Rev. 2021;31:643–82. [DOI] [PubMed] [Google Scholar]
- [16].de Almeida Macêdo E, Appenzeller S, Lavras Costallat LT. Assessment of the hospital anxiety and depression scale (HADS) performance for the diagnosis of anxiety in patients with systemic lupus erythematosus. Rheumatol Int. 2017;37:1999–2004. [DOI] [PubMed] [Google Scholar]
- [17].Wang Y, Wang Z. Effects and safety of monoamine oxidase-b inhibitors for early Parkinson’s disease: a network meta-analysis. Eur Neurol. 2024;87:273–90. [DOI] [PubMed] [Google Scholar]
- [18].Haapiainen H, Valli J, Murtola T, Huhtala H, Kaipia A, Raitanen M. Visual analogue scale (VAS) in the evaluation of functional outcomes after three-dimensional laparoscopic prostatectomy. Urology. 2023;172:121–5. [DOI] [PubMed] [Google Scholar]
- [19].Cardenas-Saenz O, Macias-Ortiz FG, Becerra-Aguiar NN, et al. Unraveling the link between sleep disturbances and Parkinson’s disease: a narrative review. Acta Neurol Taiwan. 2025;34:196–209. [DOI] [PubMed] [Google Scholar]
- [20].Huang J, Cheng Y, Li C, Shang H. Genetic heterogeneity on sleep disorders in Parkinson’s disease: a systematic review and meta-analysis. Transl Neurodegeneration. 2022;11:21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Cattaneo C, Pagonabarraga J. Sex differences in Parkinson’s disease: a narrative review. Neurol Ther. 2025;14:57–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Liguori C, Frontani F, Francescangeli G, et al. Assessment of psychometric characteristics of parkinson’s disease sleep scale 2 and analysis of a cut-off score for detecting insomnia in italian patients with Parkinson’s disease: a validation study. J Personalized Med. 2024;14:298. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Deng P, Xu K, Zhou X, et al. Constructing prediction models for excessive daytime sleepiness by nomogram and machine learning: a large Chinese multicenter cohort study. Front Aging Neurosci. 2022;14:938071. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Brusa L, Ponzo V, Stefani A, et al. Extended release levodopa at bedtime as a treatment for nocturiain Parkinson’s disease: an open label study. J Neurol Sci. 2020;410:116625. [DOI] [PubMed] [Google Scholar]
- [25].Christensen JAE, Jennum P, Koch H, et al. Sleep stability and transitions in patients with idiopathic REM sleep behavior disorder and patients with Parkinson’s disease. Clin Neurophysiol. 2016;127:537–43. [DOI] [PubMed] [Google Scholar]
- [26].Knezevic E, Nenic K, Milanovic V, Knezevic NN. The role of cortisol in chronic stress, neurodegenerative diseases, and psychological disorders. Cells. 2023;12:2726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Wang J, Sun J, Gao L, Zhang D, Chen L, Wu T. Common and unique dysconnectivity profiles of dorsal and median raphe in Parkinson’s disease. Hum Brain Mapp. 2023;44:1070–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Chen L. Effectiveness of the Roy adaptation model-based nursing intervention in improving physiological, psychological, and social outcomes in patients with Parkinson’s disease. BMC Neurol. 2025;25:219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Bohnen NI, Yarnall AJ, Weil RS, et al. Cholinergic system changes in Parkinson’s disease: emerging therapeutic approaches. Lancet Neurol. 2022;21:381–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Nyatega CO, Qiang L, Adamu MJ, Kawuwa HB. Gray matter, white matter and cerebrospinal fluid abnormalities in Parkinson’s disease: a voxel-based morphometry study. Front Psychiatry. 2022;13:1027907. [DOI] [PMC free article] [PubMed] [Google Scholar]