Depression, anxiety and somatic symptoms clustered tightly in hemodialysis care
In hemodialysis, mood symptoms rarely travel alone. This cross-sectional structural equation modeling study found that depression, anxiety, and somatic symptom burden were strongly interwoven across a large outpatient dialysis cohort, with the best-fitting model pointing to a broad internalizing construct rather than cleanly separable symptom silos.
The sample included 1,085 dialysis patients treated across 17 facilities in Northeast Ohio between 1 March 2018 and 28 February 2023. The cohort was predominantly Black, making up 75% of participants, and 150 patients, or 14.3%, screened positive for depression using a PHQ-9 score of at least 10. Those baseline numbers matter because they describe a population in which psychological distress and bodily symptom burden sit inside the same daily clinical environment as dialysis itself: fatigue, sleep disruption, pain, and treatment exhaustion can blur the boundary between psychiatric illness and physical suffering.
A single distress dimension fit the data better than isolated symptom buckets
Investigators tested single-factor, two-factor, and three-factor models before moving to bifactor and higher-order structures. All of the tested models fit reasonably well, but the three-factor model fit only slightly better than the simpler alternatives, and the inter-factor correlations were high enough to raise concern about discriminant validity. Depression and anxiety correlated at r=0.818, depression and somatic symptoms at r=0.800, and anxiety and somatic symptoms at r=0.784.
Those correlations are not trivial overlaps. They suggest that in hemodialysis patients, the symptom inventory may be capturing a broader shared state of psychological and bodily distress rather than three neatly independent diagnoses. The bifactor model performed best of the five models evaluated, with a comparative fit index of 0.957, a Tucker-Lewis fit index of 0.959, and a root mean square error of approximation of 0.043. In practical terms, the data support an overarching internalising construct that explains much of the covariance across items, while still leaving room for domain-specific contributions from depression, anxiety, and somatic complaints.
Why dialysis patients blur psychiatric and physical symptom boundaries
This pattern fits the biology and lived experience of kidney failure. Hemodialysis patients often face persistent inflammation, metabolic instability, treatment-related fatigue, disrupted sleep, reduced appetite, and the burden of repeated vascular access procedures and scheduling demands. Each of those stressors can amplify bodily discomfort and emotional distress at the same time. In such a setting, somatic symptoms are not merely background noise; they can function as a clinical bridge between mood disorder and physical illness.
The study’s modeling approach is important because standard questionnaire scores can mislead when symptoms cluster so tightly. A patient with poor energy, low appetite, and sleep disturbance may score high on both depression and somatic items even when the underlying clinical picture is more complex than a conventional diagnostic label suggests. Structural equation modeling is useful here because it tests whether the observed questions are better explained by separate latent domains or by a dominant general distress signal with smaller residual components. The latter is what emerged most convincingly.
A single distress pathway fit better than a simple three-box framework.
Demographic differences tracked with the latent distress pattern
Race, sex, and age were included as covariates in all models, and Black race, male sex, and older age were each associated with significantly different symptom patterns, although the source text truncates the exact direction and magnitude of those associations. Even with that limitation, the presence of statistically significant demographic effects signals that the latent distress construct is not distributed evenly across the dialysis population.
The finding is clinically relevant because it raises the possibility that symptom recognition and screening may vary by patient subgroup, whether because of differing symptom expression, differences in reporting, structural barriers to care, or variation in how clinicians interpret distress in dialysis settings. In a clinic where most visits are brief and dominated by lab values, fluid balance, and access concerns, mood symptoms can easily be normalized, dismissed, or folded into the expected hardship of chronic kidney disease. This study argues against that reflex.
Screening in hemodialysis may need to move beyond siloed questionnaires
The practical implication is not that PHQ-9 and anxiety screening should be abandoned. They remain useful tools. The implication is that clinicians should interpret these scores as potentially interdependent signals of a larger distress syndrome rather than independent boxes to be checked one after another. When depression, anxiety, and somatic symptoms correlate so strongly, a patient with a modestly elevated score in one domain may still merit broader assessment of sleep, pain, fatigue, coping, and functional decline.
That approach may also help explain why some dialysis patients remain symptomatic despite fragmented interventions aimed at only one domain. A patient-centered response may need to address the shared drivers of distress, including symptom burden, treatment strain, and social vulnerability, rather than assuming that a single psychiatric label captures the full picture. The bifactor structure supports that interpretation by separating the general internalizing signal from the narrower symptom domains.
The limits of cross-sectional modeling still matter at the bedside
Cross-sectional data cannot determine whether depression leads somatic symptoms, whether somatic symptoms trigger low mood and anxiety, or whether a common underlying process drives all three in parallel. The modeling results show association and structure, not temporal causation. The study also reflects one regional dialysis population in Northeast Ohio, with a high proportion of Black participants, which makes the cohort clinically valuable but also limits how broadly the estimates can be projected without caution.
Even with those constraints, the message is concrete. In hemodialysis care, distress does not appear to respect the conventional borders between psychiatry and symptom management. Depression, anxiety, and somatic complaints clustered around a shared internalizing core, and that finding should shape how dialysis teams screen, interpret, and respond to patient-reported symptoms. The unanswered question is how best to convert that latent structure into a usable bedside workflow without adding burden to already overloaded dialysis visits.