Patients newly initiating maintenance hemodialysis (MHD) face rapid changes in treatment dependence, lifestyle, symptoms, and uncertainty about disease progression. Psychological resilience—an individual’s capacity to maintain or recover psychological adaptation during stress—can influence coping, emotion regulation, quality of life, and health outcomes in MHD populations. Prior work has largely been cross-sectional or focused on stable dialysis cohorts, limiting understanding of within-person change and heterogeneous longitudinal patterns during the early dialysis transition. This study used latent growth mixture modeling to identify distinct resilience trajectories during the first six months after MHD initiation and to examine baseline sociodemographic, psychological, and clinical factors associated with trajectory membership.
This multicenter longitudinal study was conducted in four tertiary hospitals in Sichuan Province, China. Eligible participants were adults (≥18 years) who had started MHD within the previous month, could complete questionnaires, and provided written informed consent. Exclusion criteria included severe mental illness or cognitive impairment preventing questionnaire completion, severe comorbid conditions such as active malignancy, and prior peritoneal dialysis or kidney transplantation. Ethical approval was obtained from the Ethics Committee of North Sichuan Medical College.
Participants were recruited between February and July 2025 and followed for 6 months, with seven resilience assessments: baseline and monthly follow-ups for six months. Data collection was face-to-face by trained research nurses using an electronic structured questionnaire with quality controls (mandatory responses, exclusion if completion time <150 seconds). Of 892 eligible patients approached, 752 consented and completed baseline measures. To ensure sufficient longitudinal data for trajectory modeling, participants with fewer than three resilience assessments were excluded; 693 participants met this criterion and were included in the primary analyses.
Psychological resilience was measured using the Chinese Connor–Davidson Resilience Scale at seven time points. Latent growth mixture modeling (LGMM) was used to identify latent classes representing distinct longitudinal resilience patterns. Full information maximum likelihood (FIML) was applied to handle wave-level missing data under the missing-at-random assumption, allowing participants with incomplete later follow-ups to contribute available data. After LGMM class enumeration, multinomial logistic regression examined baseline factors associated with trajectory class membership.
The sample-size target considered recommendations for LGMM and the intention to compare classes on baseline characteristics. The authors sought at least 500 valid participants based on effect-size considerations for regression analyses and set a target of at least 625 to allow for ~20% attrition. The final analytic sample included 693 participants who completed at least three resilience assessments.
Baseline characteristics considered in associations with trajectory membership included sociodemographic variables (age, marital status, income), clinical factors (comorbidity burden, hospitalization frequency, physical activity), and psychological health indicators (depressive symptoms, anxiety symptoms, sleep disturbance). These variables were measured at baseline and included in multinomial logistic regression models comparing identified trajectory classes.
LGMM identified three distinct resilience trajectories among 693 patients who newly initiated MHD:
High-level stable: 51.7% of participants, showing consistently high resilience across the 6-month follow-up.
Low-stable: 35.5% of participants, exhibiting persistently low resilience over time.
High-level rapidly decreasing: 12.8% of participants, starting at a high resilience level but showing a rapid decline during early follow-up.
Compared with the high-level stable trajectory, membership in the high-level rapidly decreasing trajectory was associated at baseline with the presence of depressive symptoms, anxiety symptoms, sleep disturbance, and a higher comorbidity burden. The low-stable trajectory shared these psychological and clinical associations and was additionally associated with older age, unmarried status, lower income, more frequent hospitalizations, and lower physical activity at baseline.
These findings indicate that resilience is not uniform among new MHD patients: while about half maintain high resilience, a sizable subgroup shows persistently low resilience, and a smaller subgroup experiences early decline from initially high resilience.
To address classification uncertainty and potential attrition bias, the authors performed sensitivity analyses including the R3STEP procedure and a complete-case LGMM. These procedures were used to verify the robustness of class enumeration and baseline associations when accounting for classification error and when restricting analyses to participants with complete data. Details of numerical results from sensitivity checks were reported in the original analysis; methodological steps stated that these checks were conducted to confirm that the identified trajectory structure and associated baseline predictors were not artifacts of missingness or classification uncertainty.
This study demonstrates heterogeneous longitudinal patterns of psychological resilience during the early period after initiating MHD. Distinct trajectories had different baseline correlates: psychological distress (depression, anxiety), sleep disturbance, and greater comorbidity burden characterized both declining and persistently low resilience groups, while sociodemographic disadvantage (older age, unmarried status, lower income), higher healthcare utilization (frequent hospitalizations), and lower physical activity specifically characterized the low-stable group. The presence of a high-level but rapidly declining group highlights that an initially favorable resilience profile does not guarantee sustained adaptation during early dialysis.
Clinically, these results underscore the value of early psychosocial assessment for patients starting MHD, routine monitoring of resilience across the early months of treatment, and targeted supportive interventions for patients at risk of decline or persistent low resilience. Interventions could be prioritized for patients identified at baseline with depressive or anxiety symptoms, sleep problems, higher comorbidity burden, or sociodemographic and clinical risk factors noted above.
Among patients newly initiating maintenance hemodialysis, psychological resilience followed three heterogeneous trajectories across the first six months: a majority with stable high resilience, a substantial low-stable group, and a smaller group with rapid decline from a high baseline. Different baseline psychological, clinical, and sociodemographic factors were associated with trajectory membership. The findings support early psychosocial screening, ongoing monitoring of resilience during early dialysis, and provision of targeted supportive care for patients at risk for declining or persistently low resilience.