Construction and validation of a predictive model for depression in elderly maintenance hemodialysis patients: a multicenter study

doi:10.21203/rs.3.rs-4905140/v1

Download PDF

Research Article

Construction and validation of a predictive model for depression in elderly maintenance hemodialysis patients: a multicenter study

https://doi.org/10.21203/rs.3.rs-4905140/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Depression is the most common adverse psychological state among elderly patients undergoing maintenance hemodialysis. Persistent depressive symptoms significantly impair the patients' quality of life and increase their risk of readmission and mortality.

Method

In this study, we selected 871 elderly hemodialysis patients from the blood purification centers of nine tertiary hospitals in Chengdu, Sichuan Province, China, during the period from November 2023 to February 2024. These patients were categorized into a depression group (538 cases) and a non-depression group (333 cases). Multivariable logistic regression analysis was employed to identify independent risk factors and to develop a risk prediction model, which included the construction of a nomogram and subsequent internal validation. Subsequently, from March to April 2024, external validation of the model was performed using 219 elderly patients undergoing maintenance hemodialysis from three additional hospitals in Chengdu.

Result

The depression prevalence rate among elderly patients undergoing maintenance hemodialysis was found to be 38.2%. Logistic regression analysis revealed that education level, visual impairment, frailty, cognitive impairment, malnutrition, activities of daily living, and social support were independent risk factors for depression (p<0.05). Both internal and external validation of the model demonstrated a receiver operating characteristic curve(ROC）area under the curve (AUC) greater than 0.80, indicating good discriminative ability. Furthermore, calibration plots, the Hosmer-Lemeshow test, and clinical decision curves all demonstrated that the model has good calibration and clinical applicability.

Conclusion

In this study, we showed that the prevalence of depression is notably high among elderly patients undergoing maintenance hemodialysis. Risk factors for depression in this population include education level, visual impairment, frailty, cognitive impairment, malnutrition, activities of daily living, and social support. The nomogram prediction model developed based on these risk factors demonstrates good predictive efficacy, offering a valuable reference for healthcare professionals to identify at-risk individuals early in clinical practice.

Depression

Hemodialysis

Prediction model

Risk factors.

Maintenance hemodialysis (MHD) is the primary treatment for end-stage renal disease. With the increasing trend of population aging and the continual advancement of dialysis technology, the proportion of MHD patients aged over 60 has approached 70%, making it the fastest-growing dialysis population. Although MHD can substitute for some renal function and extend patient survival, elderly patients often face multiple complications, endure significant economic burdens, and experience considerable psychological stress. Among these complications, depression is the most common adverse psychological state in elderly MHD patients, with a prevalence rate ranging from 15–52%. Its main manifestations include low mood, loss of interest, and diminished energy, with severe cases potentially leading to suicidal tendencies [1]. Persistent depression not only reduces the quality of life for patients but may also affect the adequacy of dialysis treatment, thereby increasing the risk of readmission and mortality. Research indicates [2] that there is a close correlation between the incidence of depression and mortality in elderly MHD patients, making depression a significant predictor of patient mortality. Therefore, early identification of depression risk and timely intervention are critical for preventing the deterioration of conditions and improving adverse outcomes in elderly MHD patients. However, current studies primarily focus on exploring the influencing factors of depression, and there is a lack of specific and highly sensitive screening tools to guide healthcare professionals in early assessments of depression in elderly MHD patients. Based on this, our study employed binary logistic regression methods to identify predictive variables closely associated with the occurrence of depression and to construct a prediction model for depression in elderly MHD patients, presented in the form of a nomogram. Furthermore, to promote the widespread application of the nomogram in clinical practice, we transformed it into an easy-to-use web-based calculator, enabling healthcare professionals to quickly and accurately assess the depression risk of elderly MHD patients, thereby providing a convenient tool and method for prevention and early intervention.

2.1Study design and participants

The study employed a convenience sampling method, select 1,090 elderly patients from 12 hemodialysis centers in Sichuan Province between November 1, 2023, and April 30, 2024. The data were randomly divided into a training set and a validation set in an 8:2 ratio based on the chronological order of data collection. Specifically, 871 case data collected from November 1, 2023, to February 29, 2024, were designated as the training set, while 219 case data collected from March 1, 2024, to April 30, 2024, were allocated to the validation set. The patient recruitment process is illustrated in Fig. 1.

Inclusion criteria: (1) age ≥ 60 years; (2) patients diagnosed with end-stage renal disease (ESRD) in accordance with the guidelines established by the Kidney Disease Outcome Quality Initiative (KDOQI) [3]and the diagnostic criteria outlined in the "Guidelines for Screening, Diagnosis, and Prevention of Chronic Kidney Disease"[4], and who had received regular hemodialysis treatment for at least 3 months.

Exclusion criteria: (1) serious comorbidities such as severe infections, organ failure, or malignant tumors; (2) neurological disorders; (3) language or communication barriers that hinder understanding or response to questions. Ethical approval for this study was obtained from the Ethics Committee of Chengdu Medical College (Ethics approval number: CMCEC 2023. NO.122), and the study strictly adhered to the principles outlined in the Declaration of Helsinki. All participants provided informed consent.

2.2 Research tools

2.2.1. Study variables

In this study, the variables examined were collected using a general demographic information questionnaire developed by the researchers. This questionnaire encompasses general patient information, including age, gender, education level, marital status, body mass index, alcohol history, smoking history and duration of dialysis. Furthermore, it includes treatment factors such as visual impairment, hearing impairment, complications, as well as laboratory examination indicators including total protein, hemoglobin (Hb), albumin (Alb), high-sensitivity C-reactive protein, triglycerides (TG), total cholesterol (TC), urea, glomerular filtration rate (GFR), serum creatinine, potassium (K), calcium (Ca), sodium (Na), and phosphorus (P).

2.2.2 Depressive symptom

The 15-Item Geriatric Depression Scale (GDS-15) is employed to assess the depressive status of patients. It was developed by Sheikh and Yesavage [5] to specifically address the characteristics of the elderly, primarily evaluating symptoms such as low mood, reduced activity, irritability, withdrawal, and negative evaluations concerning past, present, and future experiences. The scale consists of 15 items, with a total score ranging from 0 to 15. Patients are instructed to respond with"yes" or "no"to each item, scoring 1 point for "yes" and 0 points for "no". A total score of ≥ 8 indicates the presence of depressive symptoms, with higher scores reflecting more severe depressive symptoms. The Cronbach's alpha coefficient for this scale is 0.82 [6, 7].

2.2.3Frailty Assessment

The Frailty Phenotype Scale was utilized to evaluate the frailty status of patients. It was developed by Fried et al. in 2001 and includes five indicators: weight loss, slowed walking speed, low grip strength, low physical activity, and fatigue. The total score ranges from 0 to 5, with scores of 3 to 5 indicating frailty, scores of 1 to 2 indicating pre-frailty, and a score of 0 indicating no frailty. This scale has been widely applied in assessing frailty among patients undergoing hemodialysis. The Cronbach's alpha coefficient for this scale is 0.93 [8, 9].

2.2.4 Cognitive function assessment

The Montreal Cognitive Assessment (MoCA) Scale [10] is utilized to evaluate eight cognitive domains, including abstract thinking, attention and concentration, executive function, language, memory, visuospatial skills, calculation, and orientation. The total score ranges from 0 to 30, with higher scores indicating better cognitive function; a score of 26 or above is considered indicative of normal cognitive function. For participants with less than 12 years of education, 1 point is added to the total score to correct for educational bias. The Cronbach's alpha coefficient for this scale is 0.81 [11].

2.2.5 Activities of Daily Living

The Activities of Daily Living Scale, developed by Mahoney [12], is employed to assess patients' ability to perform daily activities, including bowel control, bladder control, grooming, toileting, eating, bathing, dressing, transfers from bed to chair, ambulation on level surfaces, and climbing stairs, comprising a total of 10 items. The scoring ranges from 0 to 100, with a score of 40 indicating severe dependence, requiring complete care from others; scores of 41 to 60 indicating moderate dependence, necessitating most care from others; scores of 61 to 99 indicating mild dependence, requiring some assistance; and a score of 100 indicating independence, with the individual fully capable of self-care. The Cronbach's alpha coefficient for this scale is 0.88 [13].

2.2.6 Sleep Quality

The Pittsburgh Sleep Quality Index (PSQI) Scale is utilized to assess patients' sleep quality. Developed by Buysse [14]in 1989, it encompasses seven components: subjective sleep quality, sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction, totaling 18 items. Scores range from 0 to 21, with a score of 7 or higher indicating the presence of sleep problems. The Cronbach's alpha coefficient for this scale is 0.7962 [15].

2.2.7 Nutritional Status

The Malnutrition Inflammation Score (MIS) Scale, developed by Kalantar-Zadeh[16], is specifically designed to assess the nutritional status of patients with chronic kidney disease undergoing dialysis. The scale consists of four components: medical history, physical examination, body mass index (BMI), and laboratory tests, comprising a total of 10 items. Each item is scored from 0 to 3, resulting in a total score ranging from 0 to 30, with higher scores indicating a greater risk of malnutrition [17].

2.2.8 Social Support

The Social Support Rating Scale (SSRS) was compiled by Chinese scholar Xiao Shui yuan in 1994, assesses social support across three dimensions: objective support, subjective support, and utilization of social support. The scale comprises a total of 10 items, with scores ranging from 12 to 66. Higher scores indicate a greater level of social support, with a score of 22 denoting low support, 23 to 44 indicating moderate support, and 45 or above representing high support. The internal consistency of the scale, measured by Cronbach's alpha, ranges from 0.825 to 0.896 [18].

2.3Data collection

General Information and Variables Related to Depression and Malnutrition: Prior

to data collection, researchers underwent standardized training and conducted face-to-face interviews using paper-based questionnaires to retrospectively gather relevant data. All questionnaires were distributed, completed, and collected on-site.

(2) Laboratory Indicators and Disease-Related Information: Researchers utilized the Hospital Information System (HIS) to collect disease-related information and laboratory indicators for the patients, with laboratory indicators selected from the most recent test results within one month prior to enrollment. After data extraction, a double-check and cross-verification process will be conducted.

2.4 Statistical analysis

2.4.1 Binary Logistic regression analysis screened the predictor variables

SPSS 26.0 was used to perform statistical analysis of the data. Categorical variables are expressed as frequency (percentage), while continuous variables are presented as mean ± standard deviation or median (interquartile range). The Mann-Whitney U test was used for continuous variables that did not follow a normal distribution, while independent samples t-tests were applied to normally distributed variables. Chi-square tests or Fisher's exact probability method were utilized for comparing categorical variables. Independent variables with statistical significance (p < 0.05) from univariate analysis were included in the binary logistic regression analysis, employing a forward stepwise regression approach to identify the variables ultimately incorporated into the prediction model.

2.4.2 Model construction

Based on the selected predictive variables and their corresponding partial regression coefficients, a predictive model for depression in elderly patients undergoing MHD was developed. Subsequently, the nomogram was created using R software package "rms" to visually represent the risk for patients. Additionally, the "Dyn Nom" and "shiny" packages were utilized to construct and publish dynamic nomograms for falls in elderly MHD patients, with the links uploaded to a relevant website, transforming them into user-friendly web-based calculators for clinical use.

2.5Comprehensive evaluation of the model

2.5.1Model evaluation

The discrimination ability of the model refers to its capacity to correctly differentiate between different categories within the sample. The receiver operator characteristic (ROC) curve was used to evaluate the predictive ability of the model, and the area under the curve (AUC) was computed to evaluate model performance. The closer the AUC is to 1 (AUC value > 0.75), the better the discriminative validity of the model [19]. The calibration was constructed to examine the agreement between the predicted probabilities with the observed outcome, and a calibration curve was constructed and a Hosmer-Lemeshow test (H-L Test)was performed to evaluate the calibration of the nomogram. A calibration curve that closely aligns with the reference line indicates better calibration quality. In the H-L test, a P value greater than 0.05 suggests adequate calibration capability. Furthermore, a decision curve analysis (DCA) was employed to determine the clinical net benefit of the model at various threshold probabilities, thereby evaluating its clinical utility in elderly MHD patients. DCA curves were generated using the "rmda" package in R software. If the curve corresponding to a given threshold probability lies above the "None" and "All" lines, this indicates good clinical applicability of the model. To comprehensively evaluate the model's predictive performance, we also calculated sensitivity, specificity, accuracy, F1 score, and positive/negative likelihood ratios (LR+/LR-).

2.5.2Internal validation

Internal validation relies on the data used for the development of the predictive model and aims to assess the reproducibility of the model development process while guarding against overfitting that may lead to an overestimation of model performance. In this study, the Bootstrap method was employed for internal validation by performing 1,000 resampling iterations of the risk prediction model.

2.5.3 External validation

External validation assesses the performance of the model on a new dataset that was not involved in the model development process, allowing for an evaluation of the model's generalizability and predictive capability. In this study, the external validation of the model's predictive performance was evaluated using the area under the ROC curve, the Hosmer-Lemeshow goodness-of-fit test, calibration plots, and DCA.

3.1 Characteristics of the participants

After a stringent selection process based on inclusion and exclusion criteria, this study enrolled a total of 1,090 elderly patients receiving MHD. The modeling group included 871 patients, of whom 333 were diagnosed with depression (38.2%). The validation group comprised 219 patients, with 82 diagnosed with depression (37.4%). Comparisons of general characteristics between the modeling and validation groups revealed no statistically significant differences (p > 0.05). The general characteristics of the dataset are detailed in Table 1.

Table 1

Characteristics of the study population classified by depression status.
Factors	Description	Depression (n = 538)	Non-depressed (n = 333)	Test statisic	P
Sex				8.038	0.005
	male	332(61.7)	173(52)
	Female	206(38.3)	160(48)
Age (years)		68(62,74)	72(66,78)	-6.243	< .001
Duration of dialysis(month)		28(12,59)	36(13.5,60)	-2.013	0.044
BMI(kg/m²)		22.5(20.4,24.8)	22(19.7,24.2)	-2.522	0.012
Number of chronic diseases		2(1,2)	2(1,2)	-1.007	0.314
Educational level				16.573	< .001
	Elementary school or below	341(63.4)	255(76.6)
	High school and above	197(36.6)	78(23.4)
Marital status				21.457	< .001
	Married	483(89.8)	261(78.4)
	Divorced/ Widowed	55(10.2)	72(21.6)
Living situation				5.198	0.023
	Not living alone	507(94.2)	300(90.1)
	Living alone	31(5.8)	33(9.9)
Monthly household income				4.854	0.183
	< 1000	58(10.8)	34(10.2)
	1000–3000	175(32.5)	128(38.4)
	3001–5000	185(34.4)	114(34.2)
	> 5000	120(22.3)	57(17.1)
Current smoking				2.34	0.126
	Yes	72(13.4)	33(9.9)
	No	466(86.6)	300(90.1)
Current alcohol consumption				0.617	0.432
	Yes	29(5.4)	14(4.2)
	No	509(94.6)	319(95.8)
Visual impairment				44.443	< .001
	Yes	161(29.9)	175(52.6)
	No	377(70.1)	158(47.4)
hearing impairment				3.282	0.070
	Yes	118(21.9)	91(27.3)
	No	420(78.1)	242(72.7)
physical exercise				42.726	< .001
	Yes	357(66.4)	146(43.8)
	No	181(33.6)	187(56.2)
frailty				126.422	< .001
	Yes	126(23.4)	210(63.1)
	No	412(76.6)	123(36.9)
cognitive impairment				67.956	< .001
	Yes	357(66.4)	146(43.8)
	No	181(33.6)	187(56.2)
malnutrition		5(3,7)	8(5,11)	-10.214	< .001
activities of daily living		100(95,100)	90(70,100)	-12.861	< .001
social support		28(23,33)	24(10,80)	-9.169	< .001
Hemoglobin(g/L)		111(98,121.25)	109(96,120)	-1.083	0.279
High-sensitivity C-reactive protein(mg/L)		6.7(2.52,14.46)	7.4(2.6,14.46)	-0.434	0.664
Serum Albumin(g/L)		40(37,42.3)	38.9(35.9,41.15)	-3.88	< .001
Serum Phosphorus(mmol/L)		1.63(1.28,2.00)	1.64(1.26,1.99)	-0.381	0.703
Serum Calcium(mmol/L)		2.17(2.05,2.31)	2.17(2.05,2.305)	-0.322	0.748
Serum Sodium(mmol/L)		139(137,141.48)	138.6(136.4,140.6)	-2.58	0.010
Total Cholesterol(mmol/L)		3.83(3.09,4.31)	3.53(2.92,4.23)	-2.417	0.016
Triglycerides(mmol/L)		1.675(1.09,2.59)	1.48(1.02,2.37)	-2.103	0.035
Blood Urea Nitrogen(mmol/L)		19.88(14.56,23.87)	19.88(13.9,22.9)	-1.317	0.188
Serum Creatinine(mmol/L)		719(487,941)	664.9(443.8,875)	-2.639	0.008

3.2 Logistic univariate and multivariate regression analysis results

Using the presence of depression in elderly MHD patients as the dependent variable (no = 0, yes = 1), multifactorial binary logistic regression analysis was conducted with statistically significant variables from univariate analysis as independent variables. The results indicated that education level, visual impairment, frailty, cognitive impairment, malnutrition, activities of daily living (ADL) independence, and social support are independent risk factors for depression in elderly MHD patients (p < 0.05), as shown in Table 2.

Table 2

Logistic multivariate regression analysis of depression in patients.
Variables	β	SE	Wald χ2	OR	95%CI	p Value
Educational level(High school and above)	-0.47	0.197	5.712	0.625	0.425–0.919	0.017
Visual impairment(yes)	0.617	0.18	11.756	1.854	1.303–2.638	0.001
Frailty(yes)	0.906	0.193	21.99	2.475	1.694–3.614	< .001
Cognitive impairment(yes)	1.017	0.196	26.835	2.766	1.882–4.064	< .001
Malnutrition	0.091	0.031	8.765	1.095	1.031–1.162	0.003
Activities of daily living	-0.042	0.008	25.955	0.959	0.943–0.974	< .001
Social support	-0.076	0.015	27.156	0.927	0.901–0.954	< .001
Constant	5.439	1.888	8.295	230.132	——	0.004

3.3 Construction of a Depression Risk Prediction Model for Elderly MHD Patients

In this study, a prediction model was established based on seven independent risk factors for depression in elderly MHD patients: education level, visual impairment, frailty, cognitive impairment, malnutrition, activities of daily living independence, and social support, as illustrated in Fig. 2. To further simplify the calculation process, we created a web-based depression risk calculator for elderly MHD patients using the "Dyn Nom" package in R version 4.3.1, which estimates the probability of depression in this patient population. The web address is as follows: https://depressionmhd.shinyapps.io/My_DynNomapp/, with an example shown in Fig. 3.

3.4 Validation of the Depression Risk Prediction Model for Elderly MHD Patients

3.4.1 Internal validation of the modeling group

The Bootstrap method was used for 1,000 resampling iterations to evaluate the model's discrimination and calibration, with an AUC of 0.836 (95% CI: 0.809–0.864), as shown in Fig. 4A. The model exhibited a sensitivity of 0.610 and a specificity of 0.871, with detailed performance metrics presented in Table 3. The calibration curve demonstrated a close alignment with the ideal curve, yielding a mean absolute error of 0.022, as illustrated in Fig. 4B. The Hosmer-Lemeshow test resulted in (X² = 11.753, p = 0.163) (p > 0.05), indicating that the nomogram provided a good fit to the data.

3.4.2 External validation

External validation was conducted using a validation group of 219 patients, resulting in an AUC of 0.806 (95% CI: 0.746–0.866), as shown in Fig. 5A. The calibration plot indicated that the calibration curve slightly deviated from the reference line, but the overall trend was consistent, with a mean absolute error of 0.017, as illustrated in Fig. 5B. The Hosmer-Lemeshow test yielded X² = 9.702, p = 0.287. Additionally, the clinical decision curve analysis indicated that the red curves for both the modeling and validation groups were located above the None line and the All line, suggesting that the prediction model has a high net benefit and significant clinical utility, as demonstrated in Figs. 6A and 6B.

Table 3

Predictive performance of the model.
AUC (95% CI)	Sensitivity	Specificity	Accuracy	F1-score	LR+	LR-
0.836(0.809, 0.864)	0.610	0.871	0.772	0.673	4.692	0.448

The global aging population and the increase in chronic diseases have made end-stage renal disease a significant health issue, following hypertension and diabetes. While maintenance hemodialysis prolongs survival, it also brings about complications and psychological stress. Depression is prevalent among elderly dialysis patients, severely impacting their quality of life, interfering with dialysis adequacy, and increasing the risk of readmission and mortality. In our study, the incidence of depression among elderly MHD patients in the modeling group was 38.2%, while the verification group showed a rate of 37.4%. The overall incidence of depression in the sample was 38.1%, which is consistent with the findings of Zou Y (36.29%) [20], indicating that depression is a common issue in elderly MHD patients. This suggests that healthcare professionals should prioritize the assessment of depression in elderly MHD patients, enabling early detection and targeted interventions to alleviate depressive symptoms and improve prognosis.

However, due to differences in study populations and environments, existing research shows varying depression rates among elderly MHD patients. For instance, studies by Sevinc M [21] and Meng Y [22] reported rates of 42.7% and 55.1%, respectively. These discrepancies may arise from variations in sample size, geographic location, healthcare standards, and assessment tools. Additionally, the multifactorial nature of depression—including socio-demographic, psychological, disease-related, and nutritional factors—contributes to inconsistencies in reported rates. Despite these differences, there is consensus that the depression rate among elderly MHD patients is significantly higher than that of elderly community members and other patient populations, underscoring the urgency of addressing this issue. Therefore, when assessing elderly MHD patients, it is crucial to consider various factors—socio-demographic data, psychological aspects, disease-related issues, and nutritional status—to ensure accurate diagnosis and prompt intervention.

This study successfully identified independent risk factors for depression in elderly MHD patients through logistic regression analysis. The factors included education level, frailty, cognitive impairment, visual impairment, malnutrition, social support, and activities of daily living, leading to the development of a nomogram risk prediction model. Internal validation results indicated an area under the ROC curve (AUC) of 0.836 (95% CI: 0.809, 0.864), with a sensitivity of 0.610 and specificity of 0.871. External validation revealed an AUC of 0.806 (95% CI: 0.748, 0.866), demonstrating the model's strong discrimination and predictive performance. Additionally, both the internal and external validation H-L goodness-of-fit tests yielded results (p > 0.05), and calibration curves closely aligned with the ideal line, indicating good consistency of the model.A cutoff value of 0.336 was established, allowing classification of subjects into high-risk and low-risk groups, thereby providing robust support for clinical decision-making. The DCA and net benefit curves further highlighted the model's potential in optimizing patient management and resource allocation. This approach offers individualized and evidence-based depression risk estimates for elderly MHD patients, assisting clinicians in accurately identifying high-risk individuals and tailoring targeted interventions to reduce the risk of depression, ultimately improving patients' overall health and quality of life.

The results of this study indicate that educational level significantly impacts the occurrence of depression in elderly MHD patients, with an odds ratio (OR) of 0.625. This suggests that patients with a high school or vocational education or above are generally less likely to experience depression compared to those with a middle school education or lower, which is consistent with findings from Muthukumaran [23] and Ye W [24]. Patients with higher education levels tend to have better acceptance and understanding of disease diagnosis, treatment knowledge, and health decision-making, allowing them to cope more effectively with challenges during treatment and thereby reducing negative emotions. Additionally, these patients are more likely to possess self-care awareness and actively seek medical assistance, facilitating access to quality healthcare resources and information. They also demonstrate greater adaptability to life changes and social role transitions, which aids their adjustment to the hospital environment and lowers the incidence of depression. Conversely, elderly MHD patients with lower educational levels often have limited knowledge and understanding of their disease and hemodialysis, combined with a weaker ability to process complex information, leading to increased feelings of fear and anxiety, thereby elevating the risk of depression [25]. Consequently, in clinical practice, healthcare professionals should implement more detailed individualized care measures for these patients, enhancing their understanding of their condition through effective communication and guidance, and encouraging a positive mindset towards treatment and recovery.

The findings of this study reveal a significant correlation between visual impairment and the incidence of depression in elderly patients undergoing MHD. Specifically, visual impairment has been identified as an independent risk factor for depression in this population, aligning with the conclusions of Bossola [26]. Vision is a crucial sensory system for conducting normal daily activities, with approximately 80% of human information acquired visually [27]. As individuals age, visual decline becomes more pronounced, particularly among MHD patients with diabetic nephropathy, where retinal complications may exacerbate vision deterioration. Visual impairment not only hinders the daily functioning of elderly individuals but also limits their capacity to engage in work and maintain social relationships, significantly impacting both their mental and physical health as well as their overall quality of life. A meta-analysis indicates that the prevalence of depression among elderly patients with visual impairment is considerably higher than that in the general elderly population [28]. Healthcare professionals should prioritize timely vision screening and treatment for this high-risk group, integrating visual training with daily living skills and mobility exercises to address challenges related to visual impairment. Additionally, timely psychological interventions are essential for supporting mental well-being and preventing depression.

Cognitive function levels in patients are considered important predictors of depressive symptoms. In this study, cognitive impairment was identified as an independent risk factor for depression in elderly patients undergoing MHD. Elderly MHD patients with cognitive impairment have a 2.766-fold increased risk of developing depression compared to those without cognitive impairment, consistent with the findings of Feng J [29] and Jung S [30]. The prevalence of cognitive dysfunction among elderly hemodialysis patients is reported to be between 30% and 60% [31, 32]. The relationship between cognitive function and depression remains controversial, with the underlying mechanisms unclear. Some studies suggest that cognitive impairment may not be an independent risk factor, while others indicate a close association between the two [33]. Given the context of elderly MHD patients, medical practice should involve a comprehensive assessment of cognitive function and depression, exploring their interactions and shared influencing factors. A comprehensive intervention program should be developed, incorporating strategies for cognitive function enhancement and depression management to promote healthy aging.

Malnutrition has been identified as an independent risk factor for depression in elderly MHD patients. Previous research has indicated a close relationship between depression and malnutrition in this population, with a malnutrition prevalence rate as high as 52% [34]. This may be attributed to elderly individuals often experiencing insufficient nutritional intake due to tooth loss and oral degeneration, which increases the risk of malnutrition, leading to issues such as hypotension and hypoglycemia, thus affecting dialysis efficacy and resulting in loss of treatment confidence, negative emotions, and ultimately depression. Moreover, long-term dialysis treatment can result in the accumulation of middle-molecule toxins in the body, causing metabolic disorders and acid-base imbalances, which impair central nervous system function and may lead to anxiety and depression [35].

As age increases, frailty emerges as another significant issue facing elderly MHD patients and should also be recognized as a notable risk factor for depression. In this study, frailty was identified as a risk factor for depression in elderly MHD patients (OR = 2.475), consistent with the findings of Yuan H [36]. Frailty reflects a decline in physiological reserve and abnormal multi-system regulation, serving as an independent predictor of falls, disability, and mortality in older adults [37, 38]. While MHD can alleviate uremic symptoms and prolong patient life, long-term dialysis may also lead to metabolic disturbances, toxin accumulation, and physiological decline, which exacerbate the risk of frailty. Additionally, with advancing age, muscle mass and strength in elderly MHD patients gradually decrease, with a frailty prevalence of 14–37% [39]. Patients with frailty often exhibit symptoms such as reduced walking speed, decreased physical activity, and increased fatigue, which are also clinical manifestations of depression. Given the high prevalence of frailty in the elderly MHD population and its impact on depression risk, it is particularly important to enhance the assessment and screening of frailty, as well as to provide appropriate intervention guidance. Interventions such as nutrition and exercise during the pre-frailty stage can effectively improve the frail state of patients and subsequently reduce the risk of depression.

The ability to perform activities of daily living is closely related to depression in elderly patients undergoing MHD. Good self-care abilities not only help reduce the incidence of depression, consistent with Liu YM's research [40], but also promote patients in maintaining a positive mental state and social functioning. However, elderly MHD patients with diminished self-care abilities often experience a decline in independence due to prolonged illness, leading to increased dependency and a higher likelihood of negative emotions, significantly elevating the risk of depression. Additionally, a decline in daily functioning reduces social activities, exacerbating feelings of loneliness and irritability, which further deepens depressive states [41, 42]. In contrast, patients with better self-care abilities are more capable of realizing their self-worth, maintaining psychological health, and effectively alleviating depressive emotions.

Furthermore, this study found that social support is an independent predictor of depression in elderly MHD patients. Patients with low levels of social support have a higher risk of depression, which aligns with the findings of Sezer S [43] and Wang X [44]. Social support is crucial in maintaining the overall health of elderly MHD patients, as it not only enhances their confidence in social interactions but also helps them face life and illness with a more optimistic attitude, thereby preventing the onset of depressive feelings. However, due to the frequency of dialysis treatment and changes in body image, social participation among elderly MHD patients is limited, leading to reduced social functionality and decreased utilization of social support, which increases the risk of depression. Therefore, healthcare professionals need to enhance communication with patients, improve their social confidence, and foster a positive outlook on life. Hospitals can organize support group activities to provide psychological and social support. Additionally, the government and community should promote public volunteer services and recreational activities to increase opportunities for elderly individuals to interact, helping them shift their focus and better integrate into society, ultimately reducing the risk of depression.

Risk assessment is a primary measure for preventing the onset of depression. An ideal clinical prediction model should be both accurate and user-friendly. This study utilized multifactorial regression analysis to identify seven main risk factors affecting depression in elderly patients undergoing MHD and established a nomogram prediction model based on these key factors. This model transforms complex regression equations into intuitive and understandable graphics through visual analysis, enabling healthcare professionals to predict the probability of depression quickly and accurately in elderly MHD patients.

Based on the assessment results from the model, healthcare providers can precisely identify high-risk populations for depression and develop more effective, personalized treatment plans to reduce its occurrence. Furthermore, to promote the widespread use of the nomogram in clinical practice, this study also converted it into an easy-to-use web-based calculator. Healthcare professionals simply need to input the relevant predictive variable scores into the web calculator, and the system will automatically calculate the patient's risk probability for depression. The development of this web-based calculator not only simplifies the operational process and enhances prediction accuracy but also significantly improves the efficiency of healthcare professionals in screening and managing depression in elderly MHD patients.

This study has limitations: it uses a single-city sample, affecting generalizability; it’s cross-sectional and can't establish causality, necessitating prospective research for better understanding. Additionally, while high-sensitivity C-reactive protein and interleukin-6 were identified as potential risk factors, their exclusion due to regional differences underscores the need to include biological markers in future studies to improve prediction accuracy and treatment strategies for depression in elderly MHD patients.

This study identifies independent risk factors for depression in elderly patients undergoing MHD, including educational level, visual impairment, frailty, cognitive impairment, malnutrition, limited activities of daily living, and insufficient social support. Based on these key risk factors, we developed a nomogram prediction model that demonstrates excellent predictive performance and significant clinical utility in assessing depression risk among elderly MHD patients.

The design of this nomogram model is intuitive, visual, and easy for healthcare professionals to apply. Using this model, clinicians can quickly and accurately identify high-risk patients for depression, enabling timely intervention strategies to improve patients' mental states and significantly reduce the risk of depressive events. This not only enhances the overall quality of life for elderly MHD patients but also plays a crucial role in improving the efficiency and quality of clinical management.

MHD maintenance hemodialysis

ESRD end-stage renal disease

ROC receiver operator characteristic

AUC area under the curve

H-L Test Hosmer-Lemeshow test

DCA decision curve analysis

OR. odds ratio

Ethics approval and consent to participate

The study was approved by the Ethics Committee of Chengdu Medical College (2023NO.122), and the investigating students gave informed consent.

Consent for publication

Not applicable.

Availability of data and materials

The datasets generated during and analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

Competing interests

The authors declares that they have no competing interests.

Funding

This study was funded by Patient Safety Research Center of Zigong Academy of Medical Sciences (HZAQ-2023-01)

Authors' contributions

L.L. was responsible for the design of the comprehensive study, data analysis, and manuscript composition; Y.F. acquired the data, revised, and improved the manuscript; X.W. and Y.X.H. analyzed the data; Q.Y. and C.Y.H. interpreted the data and provided critical suggestions for this study, as well as helped to revise the manuscript.

Acknowledgements

Not applicable.

Dziubek W, Kowalska J, Kusztal M, Rogowski Ł, Gołębiowski T, Nikifur M, Szczepańska-Gieracha J, Zembroń-Łacny A, Klinger M, Woźniewski M. The Level of Anxiety and Depression in Dialysis Patients Undertaking Regular Physical Exercise Training - a Preliminary Study. Kidney Blood Press Res. 2016;41(1):86-98. doi: 10.1159/000368548. Epub 2016 Feb 15. PMID: 26872253.
Lerma A, Perez-Grovas H, Bermudez L, Peralta-Pedrero ML, Robles-García R, Lerma C. Brief Cognitive Behavioral Intervention for Depression and Anxiety Symptoms Improves Quality of Life in Chronic Hemodialysis Patients. Psychol Psychother. 2017 Mar;90(1):105-123. doi: 10.1111/papt.12098. Epub 2016 Jul 20. PMID: 27435635.
National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: Evaluation, Classification, and Stratification. Am J Kidney Dis. 2002 Feb;39(2 Suppl 1):S1-266. PMID: 11904577.
Expert Group of Shanghai Chronic Kidney Disease Early Detection and Standardized Diagnosis and Treatment Demonstration Project, Gao X, Mei C. Guidelines for Screening, Diagnosis, and Prevention of Chronic Kidney Disease. Pract Intern Med. 2017;37(01):28-34. doi: 10.19538/j.nk2017010108.
Sheikh JI, Yesavage JA. Geriatric Depression Scale (GDS): Recent Evidence and Development of a Shorter Version. Clin Gerontol. 1986;5(1-2):165-173. doi: 10.1300/J018V05N01_09.
Cruice M, Worrall L, Hickson L. Reporting on Psychological Well-Being of Older Adults with Chronic Aphasia in the Context of Unaffected Peers. Disabil Rehabil. 2011;33(3):219-228. doi: 10.3109/09638288.2010.503835. Epub 2010 Jul 14. PMID: 20629580.
Tang D. The Use of the Short Version of the Geriatric Depression Scale (GDS-15) in Chinese Elderly Individuals. J Clin Psychol. 2013;21(03):402-405. doi: 10.16128/j.cnki.1005-3611.2013.03.036.
Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, Seeman T, Tracy R, Kop WJ, Burke G, McBurnie MA; Cardiovascular Health Study Collaborative Research Group. Frailty in Older Adults: Evidence for a Phenotype. J Gerontol A Biol Sci Med Sci. 2001 Mar;56(3):M146-156. doi: 10.1093/gerona/56.3.m146. PMID: 11253156.
Zanotto T, Mercer TH, van der Linden ML, Koufaki P. Screening Tools to Expedite Assessment of Frailty in People Receiving Hemodialysis: A Diagnostic Accuracy Study. BMC Geriatr. 2021 Jul 2;21(1):411. doi: 10.1186/s12877-021-02356-x. PMID: 34215211; PMCID: PMC8252257.
Nasreddine ZS, Phillips NA, Bédirian V, Charbonneau S, Whitehead V, Collin I, Cummings JL, Chertkow H. The Montreal Cognitive Assessment, MoCA: A Brief Screening Tool for Mild Cognitive Impairment. J Am Geriatr Soc. 2005 Apr;53(4):695-699. doi: 10.1111/j.1532-5415.2005.53221.x.
Wang W, Wang L. Application of the Montreal Cognitive Assessment Scale in Screening Mild Cognitive Impairment Patients. Chin J Intern Med. 2007;46(05):414-416.
Mahoney FI, Barthel DW. Functional Evaluation: The Barthel Index. Md State Med J. 1965 Feb;14:61-65. PMID: 14258950.
Hou D, Zhang Y, Wu J, et al. Study on the Reliability and Validity of the Chinese Version of the Barthel Index. Clin Metaanal. 2012;27(03):219-221.
Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: A New Instrument for Psychiatric Practice and Research. Psychiatry Res. 1989 May;28(2):193-213. doi: 10.1016/0165-1781(89)90047-4. PMID: 2748771.
Taoying L, Yan L, Ping X, et al. Reliability and Validity Analysis of the Pittsburgh Sleep Quality Index. Chongqing Med. 2014;43(03):260-263.
Kalantar-Zadeh K, Kopple JD, Block G, Humphreys MH. A Malnutrition-Inflammation Score is Correlated with Morbidity and Mortality in Maintenance Hemodialysis Patients. Am J Kidney Dis. 2001 Dec;38(6):1251-1263. doi: 10.1053/ajkd.2001.29222. PMID: 11728958.
Wright M, Southcott E, MacLaughlin H, Wineberg S. Clinical Practice Guideline on Undernutrition in Chronic Kidney Disease. BMC Nephrol. 2019 Oct 16;20(1):370. doi: 10.1186/s12882-019-1530-8. PMID: 31619185; PMCID: PMC6796390.
Xiao S. The Theoretical Basis and Research Application of the Social Support Rating Scale. J Clin Psychiatr. 1994(02):98-100.
Ma H, Bandos AI, Rockette HE, Gur D. On use of partial area under the ROC curve for evaluation of diagnostic performance. Stat Med. 2013 Sep 10;32(20):3449-58. doi: 10.1002/sim.5777. Epub 2013 Mar 18. PMID: 23508757; PMCID: PMC3744586.
Zou Y, Wen Y, Liu C, et al. Anxiety and Depression in Maintenance Hemodialysis Patients and Their Relationship with Quality of Life. Chinese J Blood Purif. 2018;17(11):743-747.
Sevinc M, Hasbal NB, Sakaci T, Basturk T, Ahbap E, Ortaboz M, Mazi EE, Pirdogan E, Ling J, Unsal A. Frequency of depressive symptoms in Syrian refugees and Turkish maintenance hemodialysis patients during COVID-19 pandemic. PLoS One. 2021 Jan 4;16(1):e0244347. doi: 10.1371/journal.pone.0244347. PMID: 33395428; PMCID: PMC7781368.
Meng Y, Wu HT, Niu JL, Zhang Y, Qin H, Huang LL, Zhang XJ, Yu L, Yu HY, Yan T, Zhao JR. Prevalence of depression and anxiety and their predictors among patients undergoing maintenance hemodialysis in Northern China: a cross-sectional study. Ren Fail. 2022 Dec;44(1):933-944. doi: 10.1080/0886022X.2022.2077761. PMID: 35618386; PMCID: PMC9154798.
Aarthi Muthukumaran; Gopalakrishnan Natarajan; Dineshkumar Thanigachalam; Sheik Alavudeen Sultan; Dhanapriya Jeyachandran; Sakthirajan Ramanathan.The Role of Psychosocial Factors in Depression and Mortality Among Urban Hemodialysis Patients[J].Kidney International Reports,2021,Vol.6(5): 1437-1443.
Ye W, Wang L, Wang Y, Wang C, Zeng J. Depression and anxiety symptoms among patients receiving maintenance hemodialysis: a single center cross-sectional study. BMC Nephrol. 2022 Dec 31;23(1):417. doi: 10.1186/s12882-022-03051-8. PMID: 36585621; PMCID: PMC9804950.
Luo M, Qin X, Cen J, et al. Analysis of Depression and Its Influencing Factors in Maintenance Hemodialysis Patients. J Yuxi Nationality Medical College. 2023;45(01):109-115.
Foster R , Walker S , Brar R ,et al.Cognitive Impairment in Advanced Chronic Kidney Disease: The Canadian Frailty Observation and Interventions Trial[J].American Journal of Nephrology, 2016:473-480.DOI:10.1159/000450837.
Fu, L. X., & Wang, A. P. (2022). Research progress on self-care ability of elderly individuals with visual impairment. Evidence-Based Nursing, 8(08), 1046-1050.
Feng, S., Wu, W. D., Lu, J. L., et al. (2016). Meta-analysis of the prevalence of depression in blind and visually impaired elderly individuals. Fudan University Journal (Medical Edition), 43(02), 183-187.
Feng J, Lu X, Wang S, Li H. The assessment of cognitive impairment in maintenance hemodialysis patients and the relationship between cognitive impairment and depressive symptoms. Semin Dial. 2022 Nov;35(6):504-510. doi: 10.1111/sdi.13031. Epub 2021 Nov 2. PMID: 34726291.
Jung S, Lee YK, Choi SR, Hwang SH, Noh JW. Relationship between cognitive impairment and depression in dialysis patients. Yonsei Med J. 2013 Nov;54(6):1447-53. doi: 10.3349/ymj.2013.54.6.1447. PMID: 24142650; PMCID: PMC3809877.
Luo Y, Murray AM, Guo YD, Tian R, Ye PP, Li X, Li GG, Lu FP, Ma YC, Sun Y, Wang YZ, Xiao YF, Zhang QM, Zhao XF, Zhao HD, Chen XM. Cognitive impairment and associated risk factors in older adult hemodialysis patients: a cross-sectional survey. Sci Rep. 2020 Jul 27;10(1):12542. doi: 10.1038/s41598-020-69482-1. PMID: 32719428; PMCID: PMC7385128.
Karakizlis H, Bohl K, Ziemek J, Dodel R, Hoyer J. Assessment of cognitive impairment and related risk factors in hemodialysis patients. J Nephrol. 2022 Apr;35(3):931-942. doi: 10.1007/s40620-021-01170-3. Epub 2021 Oct 16. PMID: 34655416; PMCID: PMC8995241.
Xie, Z.W., Jin, W.D., Wang, H., et al. The impact of computerized cognitive behavioral therapy on depression and anxiety in young patients with mild cognitive impairment. Journal of Wuhan University (Medical Sciences), 2022, 43(03): 457-460. DOI:10.14188/j.1671-8852.2021.0298.
Zhu J, Chen Y, Chen L, et al. Analysis of Nutritional Status and Influencing Factors in Elderly Patients Undergoing Maintenance Hemodialysis. [J]. Preventive Medicine, 2020, 32(03): 284-288. DOI: 10.19485/j.cnki.issn2096-5087.2020.03.017.
Hua J, Zhang Y, Xu S, et al. Correlation Analysis of Depression and Malnutrition-Inflammation-Atherosclerosis Syndrome in Elderly Patients on Maintenance Hemodialysis. [J]. Journal of Chronic Diseases, 2022, 23(11): 1661-1664+1668. DOI: 10.16440/J.CNKI.1674-8166.2022.11.16.
Yuan H, Zhang Y, Xue G, Yang Y, Yu S, Fu P. Exploring psychosocial factors associated with frailty incidence among patients undergoing maintenance hemodialysis. J Clin Nurs. 2020 May;29(9-10):1695-1703. doi: 10.1111/jocn.15225. Epub 2020 Mar 12. PMID: 32092192.
Nixon AC, Bampouras TM, Pendleton N, Woywodt A, Mitra S, Dhaygude A. Frailty and chronic kidney disease: current evidence and continuing uncertainties. Clin Kidney J. 2018 Apr;11(2):236-245. doi: 10.1093/ckj/sfx134. Epub 2017 Dec 2. PMID: 29644065; PMCID: PMC5888002.
McAdams-DeMarco MA, Law A, Salter ML, Boyarsky B, Gimenez L, Jaar BG, Walston JD, Segev DL. Frailty as a novel predictor of mortality and hospitalization in individuals of all ages undergoing hemodialysis. J Am Geriatr Soc. 2013 Jun;61(6):896-901. doi: 10.1111/jgs.12266. Epub 2013 May 27. PMID: 23711111; PMCID: PMC3938084.
Exterkate L, Slegtenhorst BR, Kelm M, Seyda M, Schuitenmaker JM, Quante M, Uehara H, El Khal A, Tullius SG. Frailty and Transplantation. Transplantation. 2016 Apr;100(4):727-33. doi: 10.1097/TP.0000000000001003. PMID: 26703348.
Liu YM, Chang HJ, Wang RH, Yang LK, Lu KC, Hou YC. Role of resilience and social support in alleviating depression in patients receiving maintenance hemodialysis. Ther Clin Risk Manag. 2018 Mar 1;14:441-451. doi: 10.2147/TCRM.S152273. PMID: 29535526; PMCID: PMC5840278.
Deng Y, Paul DR. The Relationships Between Depressive Symptoms, Functional Health Status, Physical Activity, and the Availability of Recreational Facilities: a Rural-Urban Comparison in Middle-Aged and Older Chinese Adults. Int J Behav Med. 2018 Jun;25(3):322-330. doi: 10.1007/s12529-018-9714-3. PMID: 29498014.
Wang Z, Yang H, Guo Z, Liu B, Geng S. Socio-demographic characteristics and co-occurrence of depressive symptoms with chronic diseases among older adults in China: the China longitudinal ageing social survey. BMC Psychiatry. 2019 Oct 23;19(1):310. doi: 10.1186/s12888-019-2305-2. PMID: 31646992; PMCID: PMC6813124.
Sezer S, Uyar ME, Bal Z, Tutal E, Ozdemir Acar FN. The influence of socioeconomic factors on depression in maintenance hemodialysis patients and their caregivers. Clin Nephrol. 2013 Nov;80(5):342-8. doi: 10.5414/CN107742. PMID: 24091317.
Wang X, Xia F, Wang G. Mediating effect of anxiety and depression between family function and hope in patients receiving maintenance hemodialysis: a cross-sectional study. BMC Psychol. 2023 Apr 25;11(1):130. doi: 10.1186/s40359-023-01169-4. PMID: 37098642; PMCID: PMC10127155.

No competing interests reported.

Download PDF

Editorial decision: Revision requested
14 Aug, 2024
Editor assigned by journal
13 Aug, 2024
Submission checks completed at journal
13 Aug, 2024
First submitted to journal
13 Aug, 2024

You are reading this latest preprint version

Construction and validation of a predictive model for depression in elderly maintenance hemodialysis patients: a multicenter study

Status:

Version 1

Abstract

Figures

1. Introduction

2. Methods

2.1Study design and participants

2.2 Research tools

2.2.1. Study variables

2.2.2 Depressive symptom

2.2.3Frailty Assessment

2.2.4 Cognitive function assessment

2.2.5 Activities of Daily Living

2.2.6 Sleep Quality

2.2.7 Nutritional Status

2.2.8 Social Support

2.3Data collection

2.4 Statistical analysis

2.4.1 Binary Logistic regression analysis screened the predictor variables

2.4.2 Model construction

2.5Comprehensive evaluation of the model

2.5.1Model evaluation

2.5.2Internal validation

2.5.3 External validation

3. Results

3.1 Characteristics of the participants

3.2 Logistic univariate and multivariate regression analysis results

3.3 Construction of a Depression Risk Prediction Model for Elderly MHD Patients

3.4 Validation of the Depression Risk Prediction Model for Elderly MHD Patients

3.4.1 Internal validation of the modeling group

3.4.2 External validation

4. Discussion

5. Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1