Domain-specific depression patterns among hemodialysis patients in Yemen: implications for mental-health screening in resource-limited settings Authors: Nuha Al-Aghbari, Asmaa Al-Awadi, Reham Awfan, Sarah Al-Sharafi, Aya Mareai, Safia Al-Qashaee, Bushra Al-Ghaffari Published: 2026 ABSTRACT AbstractDepressive symptoms are common among patients with renal disease and are associated with poorer outcomes. This study examined depressive symptom dimensions (somatic vs. cognitive-affective) among hemodialysis patients and aimed to determine whether socioeconomic and clinical factors were differentially associated with these domains. A cross-sectional study was conducted in two major hemodialysis centers in Sana’a city between June and September 2025. Adult patients receiving maintenance hemodialysis completed the Patient Health Questionnaire-9 (PHQ-9). Multivariable logistic regression analyses were conducted to identify factors independently associated with depressive symptoms. Of the 308 participants (56.8% male), 67.2% had PHQ-9 scores ≥10, indicating moderate-to-severe depressive symptoms. Low monthly household income was independently associated with higher odds of depressive symptoms (AOR 2.18; 95% CI 1.06–4.47; p = 0.032). Female sex (p = 0.076) and the presence of comorbidities (p = 0.052) showed borderline associations with depressive symptoms. The absence of medical insurance was associated with lower odds of depressive symptoms (AOR 0.46; 95% CI 0.26–0.81; p = 0.008), although this finding should be interpreted with caution. Low income and multimorbidity were associated with somatic symptoms, whereas smoking status and income were associated with cognitive-affective symptoms. Depressive symptoms were highly prevalent among hemodialysis patients in Sana’a, highlighting the need to integrate mental-health screening into dialysis care. Reducing financial barriers to treatment should also be prioritized in dialysis care settings in Sana’a and similar resource-limited settings.Keywords: Mental health, end-stage renal disease, comorbidity, PHQ-9, low- and middle-income countries Introduction United Nations member states agreed on the ambitious Sustainable Development Goals (SDGs) with the goal of eradicating poverty and inequality, protecting the environment, and ensuring that everyone has access to health, justice, and prosperity. However, End-Stage Renal Disease (ESRD) poses significant public health challenges; an important health-related goal is to reduce non-communicable disease (NCD)-related mortality by one-third by 2030 [1]. Approximately 850 million people worldwide are estimated to have kidney disease, with the majority living in low-income and lower-middle-income countries (LICs and LMICs), and a considerable number of these persons lack access to kidney disease diagnosis, prevention, or treatment [2]. Hemodialysis (HD) is a life-sustaining treatment; however, it is associated with substantial physical, psychological, and socioeconomic burdens. Of all these, one of the most prevalent comorbidities and undertreated conditions is depressive symptoms, as some studies report that 20-40% of patients on hemodialysis worldwide experience clinically significant depressive symptoms [3,4]. The occurrence of depressive symptoms in HD patients worldwide is 2–3 times higher than in the general community [5]. Adejumo et al. [6] conducted a comprehensive meta-analysis encompassing studies from diverse global regions and reported that approximately 27% of dialysis patients screened positive for depressive symptoms. Notably, the prevalence was higher in LMICs, likely reflecting systemic disparities in healthcare access and resource availability. In Asia, prevalence ranges from 12.1% in Japan to 41% in India, arising from cultural stigma surrounding mental health, financial hardships, and lack of access to psychiatric care [7,8]. In the Middle East, findings from Saudi Arabia and Iran indicated a prevalence of depressive symptoms between 25.8 and 62% among HD patients, and the reasons for this include the high out-of-pocket treatment costs, dependency on family members, and social isolation [9,10]. Depressive symptoms among hemodialysis (HD) patients arise from a variety of factors, including biological, psychological, and social. Low income, unemployment, and a lack of medical insurance are highly associated with depressive symptoms, since financial hardship may be associated with poorer treatment adherence and higher stress levels [8]. Clinical conditions that compel the patient to spend more time in dialysis, such as frequent sessions and vascular access difficulties, including catheter-related infections, may have negatively affected mental well-being [11]. Psychosocial factors such as a lack of family support and the stigma associated with chronic disease all contribute to helplessness [12]. According to the United Nations Office for the Coordination of Humanitarian Affairs, in war zones like Yemen, these risk factors are exacerbated by disruptions in health care, increased hunger, and internal displacement, resulting in a unique yet underappreciated catastrophe that has not been described in the literature [13]. Yemen has been suffering from one of the most severe humanitarian conditions across the globe, subject to one bitter health disaster. More than fifty percent of hospitals do not function properly, while centers for dialysis become overcrowded, understaffed, and under-resourced [14]. Moreover, Yemen’s mental health crisis is a silent but urgent challenge requiring immediate attention [15]. Despite all these barriers, there is a scarcity of research assessing depressive symptoms in Yemeni hemodialysis (HD) patients, with a significant gap in the knowledge relating to the nexus between conflict and poverty when it comes to renal care outcomes. This study was conducted in two major dialysis centers in Sana’a, the biggest dialysis facilities, catering to a significant number of people with ESRD in the city. Thus, compared to single-institution research, this study profiles a larger sample of individuals within selected dialysis centers in Sana’a. Sana’a is the capital of Yemen and it is the largest city in Yemen with an estimated population of over 3 million people and is a key referral center for specialized healthcare, especially in the current context of conflict and disruption of healthcare delivery [16,17]. Therefore, this study aimed to assess the prevalence of depressive symptoms and identify factors associated with these symptoms among hemodialysis patients in Sana’a city, Yemen, with the goal of informing evidence-based clinical and policy interventions. Furthermore, it sought to delineate the distinct dimensions of depressive symptoms-somatic and cognitive-affective to evaluate whether socioeconomic and clinical predictors exerted differential influences across these symptom domains. Material and methods Study design and population This study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for cross-sectional research (See Supplementary file, Table 1). Table 1.Characteristics of the study population*.Variablen (%)Age (years), mean ± SD43.2 ± 13.6Age group <3170 (22.7) 31–50161 (52.3) ≥5177 (25.0)Sex Male175 (56.8) Female133 (43.2)Marital Status Married232 (75.3) Single76 (24.7)Literacy Status Literate205 (66.6) Illiterate103 (33.4)Employment Status Employee78 (25.3) Unemployed230 (74.7)Monthly Income Low188 (61.0) Middle73 (23.7) High47 (15.3)Residence Rural73 (23.7) Urban235 (76.3)Living Situation Family282 (91.6) Alone26 (8.4)Medical Insurance Yes123 (40.1) No184 (59.9)Smoking Status Non-smoker197 (64.0) Former smoker85 (27.6) Current smoker26 (8.4)Number of comorbidities No comorbidity86 (27.9) One comorbidity148 (48.1) Two or more comorbidities74 (24.0)Duration of hemodialysis (dialysis vintage) 3–6 months17 (5.5) 6–24 months73 (23.7) 2–5 years113 (36.7) >5 years105 (34.1)Hemodialysis frequency/ week Two times279 (90.6) Three times29 (9.4)Type of Access Arteriovenous (AV) fistula245 (79.6) AV graft27 (8.8) Hemodialysis catheter36 (11.7)Complications No98 (31.8) Yes210 (68.2)Social support Yes257 (83.4) No51 (16.6)SD standard deviation. Characteristics of the study population*. SD standard deviation. This cross-sectional study was conducted among adult patients undergoing HD in dialysis centers of Sana’a city, Yemen, from 1st June to 1st September 2025. It included voluntary consenting males and females aged ≥ 18 years who had been receiving HD for more than three months, and provided written informed consent. Patients with acute psychotic disorders or cognitive impairment were excluded. Sample size calculation and sampling method Sample size was calculated for the estimation of a single proportion with 95% confidence and 5% absolute precision. Using a conservative estimated proportion p = 0.50, Z = 1.96 and d = 0.05 Because the sampling frame (the dialysis registries of the two participating centers) contained N = 944, we applied the finite population correction. The study included about 308 participants in the analysis. This number exceeds the minimum sample size computed; all eligible patients consenting during the data collection period were recruited. Based on the rule of at least 10 outcome events per variable, the sample size was also considered to be good enough for multivariable analysis [18]. The sample of the HD patients was obtained through convenience sampling. Because only a few dialysis centers were operating and patient access was restricted due to the conflict zone’s operational boundaries, the approach was considered appropriate for use. The study included patients who received dialysis during the period of data collection, therefore all eligible patients were included. Two dialysis centers were selected to be part of the study as they continued providing dialysis treatment for ESRD patients during the conflict, unlike other centers that were disrupted. These centers are Al-Thawra Hospital, and Science and Technology Hospital. Trained research staff were present during scheduled dialysis days and approached all adult patients who met the eligibility criteria. Data collection Data on HD patient’s demographic and HD-related characteristics, along with potential risk factors for depressive symptoms, were collected using a structured questionnaire (Supplementary Table 2) through face-to-face interviews. The questionnaire was pilot tested by three field specialists for content and face validity, as well as clarity and length, with a sample of 20 HD patients who were not included in the data analysis. The final version of the data collection questionnaire was modified in response to the input received. Patients were requested to complete the PHQ-9 questionnaire within a few hours of their post-HD sessions, when uremic symptoms were at their lowest. Although efforts were made to keep things consistent, the exact timing may have been a little different for each participant because of operational complications. This precaution has not been consistently applied in similar studies previously undertaken in the Middle and Far East, resulting in a significant overestimation of the risk of depressive symptoms in such individuals [3,19]. Demographic variables have been included in this study, such as: age, sex, marital status, residence, literacy status, and employment status. Monthly household income was categorized into three groups: low (< 40,000 Yemeni Rial (YER)), middle (40,000–80,000 YER), and high (> 80,000 YER). These thresholds approximate < 75%, 75–125%, and > 125% of the estimated national median income based on available national surveys according to Al Fidah et al. [20] methodology. Living situation was classified as either living with family or alone and medical insurance was recorded as yes or no. Furthermore, the history of smoking was divided into (current smoker, former smoker, and non-smoker). Comorbidity was defined as the presence of one or more chronic conditions in an individual at the same time and it was categorized into (no comorbidity, has one comorbidity, two or more comorbidities). Duration of HD or dialysis vintage was categorized into [(3–6 months), (6–24 months), (2–5 years), and (> 5 years)] [21]. Type of access was classified as arteriovenous (AV) fistula, AV graft, and HD catheter. Complication experience after dialysis was classified into yes and no, whereas satisfaction with dialysis treatment was categorized into satisfied and unsatisfied. Frequency of missed dialysis was assessed using this question: ‘How often do you miss dialysis sessions?’ (1. Never; 2. Rarely; 3. Sometimes; 4. Often). Social support was assessed using ‘Do you receive emotional or practical support from family/friends for dialysis?’ with response option yes or no. Outcome measurement Depressive symptoms were assessed using Patient Health Questionnaire-9 (PHQ-9). It is a self-report, standardized screening tool used to assess the severity of depressive symptoms and to identify individuals with probable depression. It is not a diagnostic instrument and does not establish a clinical diagnosis of depression. The PHQ-9 is widely used in clinical and research settings due to its simplicity, reliability, and validity [22,23]. An Arabic version of the PHQ-9 has been validated among hemodialysis patients, demonstrating good internal consistency (Cronbach’s alpha of 0.86), and has been applied in dialysis populations [24,25]. The PHQ-9 contains nine items that match the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) criteria for major depressive disorder (MDD), and additionally tests for psychosocial impairment. The PHQ-9 items were rated on a scale of 0–3 (0 = not at all, 1 = several days, 2 = more than a week, and 3 = almost every day), with values equal to or more than 10 indicating a probable MDD based on the symptom frequency over the past two weeks. The PHQ-9 total score can range from 0 to 27, whereas scores of 5–9 indicate mild depressive symptoms, 10–14 indicate moderate depressive symptoms, 15–19 indicate fairly severe depressive symptoms, and ≥20 indicate severe depressive symptoms [23]. In this study, the PHQ-9 cutoff score of 10 or above was used to identify patients with moderate-to-severe depressive symptoms since its sensitivity and specificity were well balanced at this cutoff, as previously demonstrated [23,26]. The PHQ-9 items were categorized into two subscales based on prior literature: the somatic subscale (items 3, 4, 5, and 8: sleep disturbance, fatigue, appetite changes, and psychomotor changes) and the cognitive-affective subscale (items 1, 2, 6, 7, and 9: anhedonia, depressed mood, feelings of worthlessness, concentration difficulties, and suicidal ideation) [27] (see Supplementary Table 2). This two-factor structure has been supported in previous studies, particularly in medically ill populations, where somatic symptoms may overlap with physical illness, making it useful to distinguish between somatic and cognitive-affective domains [28]. In the context of hemodialysis patients, this distinction is particularly relevant, as physical symptoms related to uremia and dialysis may inflate somatic scores, whereas cognitive-affective symptoms may better reflect underlying psychological distress. Data analysis Data were analyzed using Stata MP 64-bit, version 17.0 (StataCorp LLC, College Station, TX, USA). Continuous variables were summarized using the mean and standard deviation (SD) for normally distributed data. For categorical data, frequencies and proportions were used. Because there were no missing data for the analyzed variables, complete-case analysis was conducted. The association between the independent factors and depressive symptoms was assessed using univariate analysis with binary logistic regression, and the odds ratios (ORs) and 95% confidence intervals (CIs) were provided. Differences between study sites were examined using chi-square tests for categorical variables (see Supplementary Table 4). The PHQ-9 items representing the somatic and cognitive-affective symptom scales were summed to create the two subscales. Cronbach’s alpha was used to assess internal consistency. Subscale scores were compared between participants with and without depressive symptoms using independent-samples t-tests. Multicollinearity was assessed before fitting the final model using variance inflation factors (VIFs). All VIF values were within acceptable limits (VIF < 5), with a maximum VIF of 2.07, indicating no evidence of problematic collinearity. Detailed VIF values for all variables are provided in Supplementary Table 8. Covariates were selected based on a combined, pre-specified approach: covariates a priori identified as potential confounders (sex and age) were included regardless of bivariable significance, and other covariates with p < 0.20 on bivariable analysis were included. We then constructed a full multivariable model with the variables and employed a manual backward elimination process (retaining variables with p < 0.10) while monitoring for confounding (a > 10% change in the coefficient of predictors of interest). Multivariable logistic regression was used to identify factors independently associated with depressive symptoms, and adjusted odds ratios (AORs) and 95% confidence intervals (CIs) were reported. Additionally, a sensitivity analysis using linear regression was conducted with the continuous PHQ-9 total score to examine whether similar associations were observed (see Supplementary Table 3). Furthermore, an interaction analysis between medical insurance and income was performed to assess whether the association with depressive symptoms varied across income groups (Supplementary Table 5). To make sure that the models were comparable and to avoid model instability that can happen when you choose variables repeatedly, the same set of covariates was used in the subscale analyses (somatic and cognitive-affective). Model performance was assessed using the area under the receiver operating characteristic (ROC) curve (AUC) with 95% confidence intervals to evaluate discrimination. Calibration was assessed using the Hosmer–Lemeshow goodness-of-fit test (10 groups) and calibration plots. Results Patient characteristics The study included 308 HD patients, as shown in Table 1. The mean age was 43.2 years (SD 13.6). More than half were aged 31–50 years (52.3%). Males accounted for 56.8% of the sample. Most participants were married (75.3%), unemployed (74.7%), and literate (66.6%). Low monthly income was reported by 61.0% of patients, and 59.9% had no medical insurance. The majority were urban residents and living with family members (76.3% and 91.6%, respectively). One comorbidity was reported in 48.1% of patients, and two or more in 24.0%. Arteriovenous fistula was the most common vascular access (79.6%). Twice-weekly dialysis was performed in 90.6% of patients. Complications related to dialysis were present in 68.2%. Additional descriptive variables, including dialysis-related expenses, travel time, and treatment-related factors, are presented in the Supplementary Material Table 9. Depressive symptoms (PHQ-9 ≥ 10) were present in 67.2% of patients. Of these, 33.8% had moderate, 23.4% moderately severe, and 10.1% severe depressive symptoms, as shown in Figure 1. Distribution of depressive symptoms severity based on PHQ-9 among hemodialysis patients. Figure 2 illustrates the distribution of major comorbidities among HD patients, indicating that hypertension and diabetes were the most common, followed by cardiovascular disease and viral hepatitis. Distribution of major comorbidities among HD patients (N = 308). Bivariate associations As shown in Table 2, depressive symptoms were more frequent in females than males (75.2% vs. 61.1%, p = 0.009), in unemployed participants compared with employed participants (70.4% vs. 57.7%, p = 0.038), and in patients with low income compared with those with high income (73.4% vs. 57.5%, p = 0.015). Depressive symptoms were also more frequent among those with medical insurance compared with those without (77.2% vs. 60.9%, p = 0.003). By smoking status, the prevalence of depressive symptoms was 72.9% in former smokers, 67.5% in nonsmokers, and 46.2% in current smokers (p = 0.039). Depressive symptoms increased with comorbidity burden (57.0% with none, 69.6% with one, 74.3% with two or more, p = 0.046). No significant differences were observed by age group, residence, living situation, dialysis frequency, travel time, vascular access, complications, satisfaction, or social support. Table 2.Association between baseline characteristics and depressive symptoms.VariableNormal n (%)Depressive symptoms n (%)p-valueAge group <3122 (31.4)48 (68.6)0.554 31–5057 (35.4)104 (64.6) ≥5122 (28.6)55 (71.4) Sex Male68 (38.9)107 (61.1)0.009 Female33 (24.8)100 (75.2) Marital status Married78 (33.6)154 (66.4)0.588 Single23 (30.3)53 (69.7) Literacy status Literate71 (34.6)134 (65.4)0.331 Illiterate30 (29.1)73 (70.9) Employment status Employee33 (42.3)45 (57.7) 0.038 Unemployed68 (29.6)162 (70.4) Monthly income Low50 (26.6)138 (73.4) 0.015 Middle31 (42.5)42 (57.5) High20 (42.6)27 (57.4) Residence Rural20 (27.4)53 (72.6)0.261 Urban81 (34.5)154 (65.5) Living situation Family90 (31.9)192 (68.1)0.28 Alone11 (42.3)15 (57.7) Medical insurance Yes28 (22.8)95 (77.2) 0.003 No72 (39.1)112 (60.9) Smoking status Non-smoker64 (32.5)133 (67.5) 0.039 Former smoker23 (27.1)62 (72.9) Current smoker14 (53.9)12 (46.1) Number of comorbidities No comorbidity37 (43.0)49 (57.0) 0.046 One comorbidity45 (30.4)103 (69.6) Two or more comorbidities19 (25.7)55 (74.3) Duration of hemodialysis 3–6 months8 (47.1)9 (52.9) 0.09 6–24 months20 (27.4)53 (72.6) 2–5 years31 (27.4)82 (72.6) >5 years42 (40.0)63 (60.0) Family history of renal failure No89 (34.1)172 (65.9)0.249 Yes12 (25.5)35 (74.5) Hemodialysis frequency Two times89 (31.9)190 (68.1)0.301 Three times12 (41.4)17 (58.6) Monthly dialysis expenses <20,000 YER42 (38.2)68 (61.8)0.107 20,000–50,000 YER10 (43.5)13 (56.5) >50,000 YER49 (28.0)126 (72.0) Travel time <30 min48 (30.6)109 (69.4)0.519 30–60 min31 (37.8)51 (62.2) >1 h22 (31.9)47 (68.1) Type of access AV fistula80 (32.7)165 (67.3)0.095 AV graft13 (48.1)14 (51.9) Catheter8 (22.2)28 (77.8) Complications No37 (37.8)61 (62.2)0.205 Yes64 (30.5)146 (69.5) Satisfaction with dialysis treatment Satisfied91 (33.5)181 (66.5)0.495 Unsatisfied10 (27.8)26 (72.2) Missed sessions Never82 (31.3)180 (68.7)0.401 Rarely15 (40.5)22 (59.5) Sometimes4 (44.4)5 (55.6) Social support Yes81 (31.5)176 (68.5)0.285 No20 (39.2)31 (60.8) Data are presented as n (%). P-values were calculated using Pearson’s chi-square test and are unadjusted for multiple comparisons. Association between baseline characteristics and depressive symptoms. Data are presented as n (%). P-values were calculated using Pearson’s chi-square test and are unadjusted for multiple comparisons. Additional analysis was conducted to assess the differences between study sites, which showed that there was no significant difference in the prevalence of depressive symptoms between the two study sites (χ2 = 0.03, p = 0.855) (see Supplementary Table 4). Multivariable analysis Table 3 presents the results of the logistic regression. Low income was associated with higher odds of depressive symptoms (AOR 2.18, 95% CI 1.06–4.47, p = 0.032). Absence of insurance was associated with lower odds (AOR 0.46, 95% CI 0.26–0.81, p = 0.008). Female sex showed an elevated but non-significant odds ratio (AOR 1.70, 95% CI 0.94–3.07, p = 0.076). Patients with one comorbidity had higher odds (AOR 1.82, 95% CI 0.99–3.35, p = 0.052); two or more comorbidities (AOR 1.81, 95% CI 0.86–3.82, p = 0.114). Employment status, residence, and smoking were not significant factors associated with depressive symptoms. Table 3.Multivariable logistic regression analysis of factors associated with depressive symptoms.VariablesAOR95% CIp-valueAge [Ref: <31] 31–500.900.46 − 1.760.764 ≥511.240.55 − 2.790.600Sex [Ref: Male] Female1.700.94–3.070.076Employment Status [Ref: Employee] Unemployed1.300.70 − 2.400.401Monthly Income [Ref: High] Low2.181.06–4.47 0.032 Middle1.180.54 − 2.610.667Residence [Ref: Rural] Urban0.850.45 − 1.610.626Medical Insurance [Ref: Yes] No0.460.26–0.81 0.008 Smoking Status [Ref: Non-smoker] Former smoker1.110.57 − 2.120.752 Current smoker0.470.19 − 1.190.116Number of Comorbidities [Ref: No comorbidity] One comorbidity1.820.99–3.350.052 Two or more comorbidities1.810.86–3.820.114AOR: Adjusted odds ratio; CI: confidence interval. Multivariable logistic regression analysis of factors associated with depressive symptoms. AOR: Adjusted odds ratio; CI: confidence interval. Sensitivity analysis using linear regression with the continuous PHQ-9 score showed results consistent with the primary logistic regression model. Low income and comorbidity burden were associated with higher PHQ-9 scores, whereas absence of medical insurance was associated with lower scores. These results support the robustness of the main findings (Supplementary Table 3). Additional analysis including an interaction term between medical insurance and income showed no significant interaction (all p > 0.05), indicating that the association between insurance status and depressive symptoms was consistent across income groups (Supplementary Table 5). Furthermore, analysis showed that smoking status was significantly associated with comorbidity burden (p = 0.010), and the association between former smoking and depressive symptoms was no longer significant after adjustment for comorbidity, suggesting a potential confounding effect (Supplementary Tables 6 and 7). Model performance The final multivariable logistic regression model demonstrated acceptable discrimination, with an AUC of 0.70 (95% CI: 0.65–0.75). Calibration was good, as indicated by the Hosmer–Lemeshow test (χ2(8) = 5.40, p = 0.71), suggesting close agreement between observed and predicted risks of depressive symptoms, as shown in Figure 3. Receiver operating characteristic (ROC) curve of the final multivariable logistic regression model predicting depressive symptoms among hemodialysis patients (N = 308). The model demonstrated acceptable discrimination with an area under the curve (AUC). When comorbidity was modeled as a continuous variable, the number of comorbidities was significantly associated with depressive symptoms. Each additional comorbidity was associated with higher odds of depressive symptoms by 37% (OR = 1.37, 95% CI: 1.00–1.85, p = 0.048), suggesting a dose–response pattern. Comparison of somatic and cognitive-affective scores by depression status The somatic subscale showed acceptable internal consistency (Cronbach’s α = 0.69), while the cognitive-affective subscale’s reliability was lower (Cronbach’s α = 0.57). Participants with depressive symptoms scored significantly higher than those without depressive symptoms on both the somatic (M = 8.36 vs. 3.36; t(306) = −19.91, p < 0.001) and cognitive-affective subscales (M = 6.70 vs. 2.39; t(306) = −14.96, p < 0.001). Linear regression analysis, as shown in Table 4, indicated that low income and female sex were significantly associated with higher somatic symptom scores, whereas income and smoking status were associated with cognitive-affective symptom scores. The absence of medical insurance was associated with lower depressive symptom scores in both domains. The models explained 16% and 12% of the variance in somatic and cognitive-affective scores, respectively, indicating modest explanatory power (adjusted R2 = 0.13 and 0.09, p < 0.001). Table 4.Linear regression models predicting somatic and cognitive-affective depressive symptom subscale scores among hemodialysis patients. Somatic scale modelCognitive-Affective scale modelCharacteristicsβ (coefficient)95% CIp-valueβ (coefficient)95% CIp-valueSociodemographic CharacteristicsAge [Ref: <31] 31–500.24−0.60 − 1.090.576−0.16−1.03 − 0.710.716 ≥510.81−0.19 − 1.830.1150.07−0.97 − 1.110.895Sex [Ref: Male] Female0.960.21 − 1.70 0.011 0.07−0.68 − 0.830.853Employment Status [Ref: Employee] Unemployed0.39−0.42 − 1.200.3430.62−0.21 − 1.450.142Monthly Income [Ref: High] Low1.560.61 − 2.51 0.001 1.170.20 − 2.15 0.018 Middle0.32−0.74 − 1.400.5461.220.12 − 2.32 0.029 Residence [Ref: Rural] Urban−0.16−0.94 − 0.610.682−0.68−1.47 − 0.110.094Medical Insurance [Ref: Yes] No−1.06−1.75 − 0.37 0.003 −1.28−1.99 – −0.57 <0.001 Smoking Status [Ref: Non-smoker] Former smoker0.35−0.46 − 1.170.3990.32−0.51 − 1.170.442 Current smoker−0.40−1.63 − 0.830.523−1.49−2.76 – −0.23 0.020 Number of Comorbidities [Ref: Non comorbidity] One comorbidity0.860.06 − 1.67 0.033 0.890.08 − 1.71 0.031 Two or more comorbidities1.670.72 − 2.61 0.001 1.260.29 − 2.23 0.011 Model summary R²0.160.12 Adj R²0.130.09p-value <0.001 <0.001 Linear regression models predicting somatic and cognitive-affective depressive symptom subscale scores among hemodialysis patients. Figure 4 illustrates the mean scores for the somatic and cognitive-affective depressive symptom subscales by income and sex categories. In both subscales, the mean scores of female patients were higher than those of male patients; the most notable difference was in somatic symptoms (7.3 vs. 6.2). Mean somatic and cognitive-affective PHQ-9 subscale scores among hemodialysis patients (N = 308). Panel A shows mean scores by sex, and Panel B shows mean scores by income level. Error bars represent 95% confidence intervals. Differences between groups. Comparing income groups, it was found that patients with lower incomes had higher mean scores for both the somatic (7.3) and cognitive-affective (5.5) domains than patients with higher incomes (5.7; 4.3). These patterns suggest that some forms of depressive symptoms, particularly those that are somatic or physically manifested, are more severe when gender and low economic status are combined. Discussion As the first study in Yemen to examine the somatic and cognitive-affective dimensions of depressive symptoms among dialysis patients, this study provides novel insights into how clinical and socioeconomic factors are associated with these domains. This study found a high prevalence of depressive symptoms among HD patients in Sana’a city, with more than two-thirds of participants (67.2%) screening positive for depressive symptoms, which was substantially higher than global pooled estimates of 2(7–40%) [29], and those reported in other LMICs (25–41%) [8,9]. The elevated burden in Yemen underscores the compounding effects of socioeconomic hardship, healthcare system collapse, and ongoing armed conflict and humanitarian crisis, all of which may exacerbate psychosocial distress and limit access to adequate mental health services [13,30]. Although the prevalence identified in our study is higher than what is generally reported in the broader population, the study population differs from the general population in several significant ways, such as a greater prevalence of comorbid conditions, greater financial hardship, and a reliance on long-term dialysis care. These factors may contribute to the elevated prevalence of depressive symptoms observed in this group. In addition, the possibility of reverse causality should also be considered, as some patients may have experienced depressive symptoms prior to initiating dialysis, while others may develop these symptoms as a result of treatment burden or the advancement of their illness. Female sex was associated with a higher prevalence of depressive symptoms in bivariate analysis, consistent with prior studies suggesting increased vulnerability among women due to biological susceptibility, greater caregiving burdens, and sociocultural constraints [3]. Although the association did not reach statistical significance in multivariable analysis, the observed trend aligns with Middle Eastern studies that highlight gender-related disparities in chronic disease coping and access to psychosocial support [10]. Socioeconomic indicators emerged as important factors associated with depressive symptoms, particularly low income was associated with depressive symptoms, which aligns with prior evidence that financial strain and treatment-related expenses may act as strong psychological stressors among HD patients [12,31]. Surprisingly, lack of medical insurance was inversely associated with depressive symptoms. This finding may reflect selection or healthcare-seeking bias, whereby insured patients may have greater healthcare engagement, which may contribute to a higher likelihood of depression screening and diagnosis, or heightened awareness of their disease burden. Alternatively, given Yemen’s fragmented insurance system, insured patients may experience unmet expectations regarding coverage, which may further aggravate distress [20]. This finding should be interpreted with caution, and residual confounding factors cannot be excluded. This paradox warrants deeper qualitative exploration. Patients with one or more chronic illnesses had higher odds of depressive symptoms, with a clear dose–response relationship suggesting that multimorbidity may increase psychological vulnerability. This finding resonates with global studies showing that depression in dialysis patients is strongly associated with cardiovascular disease, diabetes, and infection-related complications, which may increase symptom burden and impair quality of life [11,32]. Interestingly, smoking status showed an unexpected pattern; former smokers had the highest prevalence of depressive symptoms. Similar patterns have been reported in other chronic disease populations and may reflect reverse causation, whereby patients with advanced disease discontinue smoking but remain psychologically distressed [33]. This pattern may also reflect greater illness severity among former smokers, as patients with multiple comorbidities or declining health are more likely to quit smoking following medical advice. Additionally, unmeasured confounding factors, such as duration and intensity of smoking, underlying disease severity, or psychosocial stressors, may have influenced this association. The attenuation of this association after adjustment for comorbidity suggests that disease severity may partly explain this finding. The distinction between somatic and cognitive-affective symptom domains provided additional insight into the nature of depressive symptoms in this population. Somatic symptoms are more common among females and low-income patients, whereas cognitive-affective symptoms were more closely related to socioeconomic and behavioral factors. This distinction is particularly relevant in HD populations, where somatic symptoms may overlap with uremic or treatment-related effects. These findings were consistent with previous studies conducted among hemodialysis and ESRD patients [34]. Moreover, findings from general or outpatient settings in LMICs, such as rural Ethiopia, suggest that emotional and cognitive symptoms may be more specific markers of clinical depression than somatic symptoms, which may be less specific due to their overlap with physical illnesses [35]. These findings support our hypothesis that, among the hemodialysis patients in our study sample in Sana’a, somatic symptoms may be more strongly associated with physical comorbidity and gender-based vulnerabilities, while cognitive-affective symptoms are better at capturing the influence of socioeconomic and psychosocial factors. The PHQ-9 items have been classified into somatic and cognitive-affective categories to provide a more thorough investigation of depressive symptoms in medically ill patients, particularly those receiving dialysis. Dialysis patients often have somatic symptoms (e.g. fatigue, changes in appetite, sleep disruption) that overlap with uremic symptoms, which raises the overall PHQ-9 scores [36]. However, the relatively low internal consistency observed for the cognitive-affective subscale (Cronbach’s α = 0.57) should be interpreted with caution. This may reflect cultural differences in the expression of psychological symptoms, overlap with physical illness manifestations, or the limited number of items included in this subscale. Additionally, the smaller number of items in this subscale may contribute to reduced reliability. Similar findings have been reported in studies of medically ill populations, where cognitive-affective symptoms may be less consistently captured than somatic symptoms. Recent research suggests that cognitive-affective symptoms, rather than somatic symptoms, have a greater influence on hospitalization and death prediction in HD populations [34]. Healthcare professionals may thus better prioritize screening and treatment in underserved regions with limited medical resources by using the two-domain technique to distinguish between depressive symptoms–related and disease-related problems. These findings highlight the importance of integrating mental health screening into routine dialysis care, particularly in resource-limited and conflict-affected settings. Targeted interventions addressing financial hardship, multimorbidity, and psychosocial support may help address the burden of depressive symptoms among HD patients. Strengths and limitations This study contributes novel data on the mental health of HD patients in Yemen, filling an important evidence gap in conflict-affected contexts. Unlike many prior single-center studies, this study drew participants from two of the largest dialysis facilities in Yemen, improving the representativeness of the findings for urban ESRD populations in Sana’a city, rather than the national population. Additionally, the use of a validated screening tool and structured interviews adds methodological rigor. Administering the PHQ-9 shortly after dialysis reduces misclassification caused by transient uremic or procedure-related somatic complaints, improving the specificity of the instrument in this setting. In addition, use of a validated Arabic PHQ-9 enhances cultural and linguistic validity; the Arabic translation used in this study has shown good internal consistency in the previous studies, which supports the reliability of our prevalence estimates and associations in this population. However, the cross-sectional design precludes causal inference, and the use of convenience sampling from only two dialysis centers in Sana’a may limit the generalizability of the findings. Therefore, the results should be interpreted with caution and may not be representative of all hemodialysis patients across Yemen, particularly those in rural areas or underserved regions. The reliance on self-reported socioeconomic and clinical data could introduce reporting bias. Furthermore, cultural differences in symptom expression may influence PHQ-9 responses, though the tool has been validated in Arabic-speaking dialysis populations. Additionally, the use of a self-reported screening tool rather than clinical diagnostic interview may lead to misclassification of depressive symptoms. Moreover, the unequal distribution of participants between the two sites may limit the power to detect site-specific differences. Finally, although the PHQ-9 was generally administered after dialysis sessions to minimize the influence of uremic symptoms, slight variation in timing may have occurred, which could have influenced symptom reporting. Policy and clinical implications The findings highlight an urgent need to integrate mental health care into dialysis care settings in Yemen, particularly in urban and dialysis centers operating in conflict-affected areas. Screening for depressive symptoms should be systematically implemented at each dialysis session using a brief tool, administered by trained nurses or dialysis staff during routine visits, coupled with counseling and low-cost psychosocial interventions adapted to resource-constrained settings. Community-based support groups and family-centered education may mitigate stigma and strengthen social support. At the policy level, expanding financial protection schemes should involve subsidizing out-of-pocket costs for dialysis and incorporating mental health services into current humanitarian health programs. Furthermore, enhancing the continuity of dialysis care necessitates the establishment of stable supply chains for dialysis consumables, the maintenance of uninterrupted treatment schedules, and the promotion of healthcare workforce retention in conflict-affected regions. Increasing humanitarian investment in mental health should focus task-shifting (e.g. training non-specialist health workers), adding mental health services to primary care and dialysis services, and developing culturally appropriate intervention programs that fit the Yemeni context. These specific strategies are essential to address the double burden of ESRD and mental health in fragile health systems. Conclusion Depressive symptoms were highly prevalent among hemodialysis patients in Sana’a, with more than two-thirds of the participants screening positive. Low income and comorbidity burden were independently associated with higher levels of depressive symptoms, while the reverse association with medical insurance should be interpreted with caution. The classification of depressive symptom domains enabled the identification of various socioeconomic and clinical factors, highlighting the need for domain-specific screening and individualized psychosocial support in dialysis care. Amid Yemen’s ongoing conflict and fragile healthcare system, these findings highlight the need for integrated, holistic approaches that coordinate renal care with universal mental health screening, counseling, and social interventions. Strengthening financial protection and lowering the cost of treatment may also alleviate psychological distress. Future studies should investigate longitudinal depressive symptoms trajectories among patients on dialysis, evaluate the effectiveness of culturally adapted psychosocial interventions, and further study the counterintuitive impact of medical insurance in this setting. Future research can also examine whether somatic and cognitive-affective symptoms follow distinct pathways or respond differently to pharmacologic or psychosocial interventions in hemodialysis patients. This research will help to identify significant gaps and provide critical evidence to inform health policy and improve survival and quality of life among ESRD survivors in areas of conflict. Supplementary Material Supplementary File.docx Acknowledgements The authors thank the doctors, nurses, and administrative staff of the dialysis centers. The authors also thank all HD patients who consented to participate in this study. N.A. also wishes to express her sincere gratitude to the Kuwaiti Women Philanthropic Team for their generous support and encouragement throughout her academic journey. Funding Statement The author(s) reported there is no funding associated with the work featured in this article. Ethics approval and consent to participate This study was conducted in accordance with the ethical principles of the Declaration of Helsinki regarding research on human subjects. The study protocol was evaluated and approved by the Research Ethics Committee of the Faculty of Medicine and Health Sciences at the University of Science and Technology in Sana’a, Yemen [Study ID: 1446/0057/UREC/UST]. Furthermore, all participants were provided written informed consent after receiving a thorough description of the study’s aims. Participants were informed that they might withdraw from the research at any moment without providing a reason. The privacy and confidentiality of participants’ data were protected. Disclosure statement No potential conflict of interest was reported by the author(s). Data availability and materials statement The current study’s datasets are accessible from the corresponding author upon reasonable request, and all relevant information is included in the manuscript. References Associated Data Supplementary Materials Supplementary File.docx Data Availability Statement The current study’s datasets are accessible from the corresponding author upon reasonable request, and all relevant information is included in the manuscript.