The association of positive and negative affect with care dependence in elderly patients with senile cataract: the mediating role of self-regulatory fatigue Authors: Xue Zhao, Shunmei Qiu, Changhong Ye, Xiaowen Sun, Yingying Luo, Yuan Liu Published: 2026 ABSTRACT Senile cataract is a leading cause of visual impairment in older adults worldwide, yet the psychological mechanisms (particularly the role of self-regulatory fatigue) through which affective states influence care dependence in this population remain poorly understood. A cross-sectional survey was conducted in the Ophthalmology Outpatient Department of the Affiliated Hospital of Guizhou Medical University from January to May 2025. Positive and negative affect were assessed using the Chinese-revised Positive and Negative Affect Schedule Short (PANAS), self-regulatory fatigue using the Self-Regulatory Fatigue Scale (SRF-S), and care dependency using the Care Dependency Scale (CDS). Correlation analyses were conducted, and a path analysis using structural equation modeling was performed to evaluate direct and indirect effects while adjusting for the visual acuity of the affected eye, age, gender, monthly per capita income, and marital status. The study included 452 participants. The parsimonious path analysis showed that positive affect was positively associated with the CDS score (standardized β = 0.324, P < 0.001; higher CDS scores indicate greater independence/lower care dependence) and negatively associated with self-regulatory fatigue (standardized β = − 0.367, P < 0.001). In contrast, negative affect was negatively associated with the CDS score (standardized β = − 0.139, P = 0.004), indicating greater care dependence, and positively associated with self-regulatory fatigue (standardized β = 0.543, P < 0.001). Self-regulatory fatigue was negatively associated with the CDS score (standardized β = − 0.184, P = 0.001), also indicating greater care dependence. Bootstrap analysis confirmed significant indirect effects of positive affect (standardized β = 0.068, 95% CI 0.028–0.109, P = 0.003) and negative affect (standardized β = − 0.099, 95% CI − 0.152 to − 0.043, P = 0.002) on the CDS score through self-regulatory fatigue. These findings highlight associations between emotional and self-regulatory states and care dependence (as reflected in the CDS score) in elderly cataract patients.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-57539-6. Introduction Senile cataract represents a major public health challenge globally, with approximately 94 million individuals aged 50 and older worldwide experiencing blindness or visual impairment due to cataract in 20201,2. Global cataract prevalence rose from 42.332 million in 1990 to 100.571 million in 2021, an increase of nearly 138%3. In China specifically, more than one-fifth (approximately 22%) of people aged 45–89 suffered from cataract between 1990 and 2015, and this figure is projected to surpass 240.83 million by 20504. Notably, 74.5% of elderly Chinese cataract patients delay necessary treatment, and 38.1% demonstrate poor disease knowledge5, highlighting the need to understand factors beyond visual acuity that shape patient outcomes. Beyond visual impairment itself, cataracts significantly compromise functional independence in older adults. Studies show that 58% of patients report mild interference with normal daily life, while 16% experience severe functional limitations5. Emotional states have garnered increasing attention as determinants of functional outcomes in this population. Positive affect, reflecting active engagement and enthusiasm, is conceptualized as a protective resource that supports adaptive coping and goal-directed behavior under stress, and has been empirically linked to lower risk of functional decline in older adults6–9. Negative affect, reflecting distress-related emotions such as upset, nervousness, and fear10, can increase perceived burden, reduce activity initiation, and interfere with self-care routines, thereby increasing care dependence11. Longitudinal evidence further suggests that depressive symptoms predict subsequent functional decline, particularly in older adults facing sensory impairment12. Despite these associations, the psychological mechanisms linking affective states to care dependence in elderly cataract patients remain insufficiently tested. Self-regulatory fatigue, a state of diminished capacity for self-control and goal-directed regulation following sustained demands, offers a theoretically grounded mediating mechanism. According to the strength model of self-control, self-control resources are finite; sustained negative affect increases regulatory demands through ongoing monitoring and suppression of distressing emotions, accelerating resource depletion and elevating self-regulatory fatigue13–15. Conversely, positive affect supports broader cognitive–behavioral repertoires and more adaptive coping strategies, which may conserve or restore regulatory capacity under stress7,8. Visual impairment further compounds this process by increasing the compensatory cognitive workload required for everyday tasks, consuming executive resources, and contributing to fatigue16,17. When self-regulatory fatigue is elevated, patients may have reduced capacity to initiate and sustain self-care behaviors, thereby increasing care needs17,18. Empirical evidence for this emotion-self-regulatory fatigue-care dependence pathway in cataract populations remains limited. Existing research has largely focused on direct associations between visual impairment and mental health outcomes, for example, demonstrating that first-eye cataract surgery reduces mental health outcomes for depression and anxiety by 18.8% and decreases depression risk by 25%19,20, but has not tested self-regulatory fatigue as a mediator of the affect-care dependence relationship19,21,22. Furthermore, evidence on this pathway is scarce in the Chinese healthcare context, where delayed care-seeking and variations in family support may uniquely shape daily functioning and care needs5. Therefore, the present study aims to examine the impact of positive and negative affect on care dependence in Chinese patients with senile cataract and to determine whether self-regulatory fatigue mediates this relationship. This study tested an affect → self-regulatory fatigue → care dependence pathway, controlling for visual acuity and key demographic factors. It was hypothesized that negative affect would be associated with increased care dependence through higher self-regulatory fatigue, whereas positive affect would be associated with lower self-regulatory fatigue and higher CDS scores, reflecting greater independence. Methods Study design and participants A cross-sectional study was conducted at the Ophthalmology Outpatient Department of the Affiliated Hospital of Guizhou Medical University between January and May 2025, involving consecutive patients diagnosed with senile cataract. Participants were eligible for inclusion if they were 60 years of age or older, had received a diagnosis of senile cataract based on the diagnostic criteria outlined in the ninth edition of Ophthalmology, and had this diagnosis confirmed by relevant clinical examinations. Only individuals without a history of cataract extraction or any other ocular surgery in either eye were considered. In addition, participants were required to exhibit best-corrected visual acuity (BCVA) of 0.5 or lower in the better-seeing eye. Informed consent was obtained from all individuals prior to enrollment, reflecting their voluntary agreement to participate in the study. Exclusion criteria were applied to individuals presenting with severe psychiatric disorders, cognitive impairment, or communication difficulties that could interfere with their ability to complete assessments or provide reliable responses. Cognitive impairment was screened using clinical judgment based on medical records and the face-to-face interview, including the participants’ ability to understand the study information, communicate consistently, and provide coherent responses; no standardized cognitive scale was administered. Patients with a current or prior history of substance dependence, including drug or alcohol abuse, were also excluded. Further grounds for exclusion included the presence of serious systemic illnesses, such as advanced cardiopulmonary or renal dysfunction, active malignancies, or autoimmune diseases. In addition, individuals with concurrent ocular conditions known to affect visual function, including glaucoma, macular degeneration, or other retinal pathologies, were not eligible to participate. Participants were subsequently excluded from the final analysis if they provided incomplete questionnaire data or withdrew informed consent at any stage of the research process. The study protocol received approval from the institutional ethics committee, under approval number [2025152 K], and was conducted in accordance with the principles outlined in the Declaration of Helsinki. Variables and scales Positive and negative affect schedule Originally developed by Watson and colleagues in 1988, the Positive and Negative Affect Schedule (PANAS) is a psychometric tool designed to assess emotional states along two distinct dimensions: positive affect and negative affect10. The scale consists of 20 items: 10 assessing positive affect and 10 assessing negative affect. Respondents rate each item on a 5-point Likert scale, indicating the extent to which they have experienced each emotion. The Chinese adaptation of the PANAS includes 9 descriptors for each of positive and negative affect. The internal consistency of the Chinese version is robust, with Cronbach’s α exceeding 0.85 for the positive affect subscale and reaching 0.84 for the negative affect subscale across two self-reported administrations. Additionally, peer-reported internal consistency is reported at 0.7723. Self-regulatory fatigue scale Self-Regulatory Fatigue Scale (SRF-S), developed by Canadian researcher Nes, is designed to measure trait-level self-regulatory fatigue, often conceptualized as self-regulatory fatigue24. The instrument comprises 16 items spanning three dimensions: cognitive, emotional, and behavioral fatigue. Five items are reverse-coded to control for response bias. Each item is rated on a 5-point Likert scale, ranging from “1 = strongly disagree” to “5 = strongly agree,” with higher cumulative scores indicating greater levels of self-regulatory fatigue. The Chinese version of the SRF-S demonstrates satisfactory internal reliability, with a Cronbach’s α coefficient of 0.8425. Care dependency scale Care Dependency Scale (CDS), developed by Dijkstra and colleagues, is intended to evaluate the extent of care dependency in individuals receiving clinical or institutional care26. The scale includes 15 items that assess physiological, psychological, and social aspects of care needs. Responses are scored on a 5-point Likert scale, yielding total scores ranging from 15 to 75. Lower scores reflect higher levels of dependency. Dependency levels are categorized as follows: complete dependence (< 25), high dependence (25–44), partial dependence (45–59), limited dependence (60–69), and complete independence (> 69). A total score of 68 or below is operationally defined as indicating care dependency. The CDS demonstrates acceptable psychometric properties, with inter-rater reliability ranging from 0.51 to 0.83 and test–retest reliability from 0.66 to 0.89. The internal consistency of the original scale is high, with a Cronbach’s α coefficient of 0.97. The Chinese version, translated and culturally adapted by domestic researchers, also demonstrates strong internal reliability (Cronbach’s α = 0.9528). Data collection procedures Data collection was implemented in two distinct phases: a preliminary pilot test and a subsequent formal investigation. The pilot test was conducted among 32 patients with senile cataract who met the predefined inclusion and exclusion criteria. Its primary aim was to assess the feasibility and internal consistency of the survey instruments. The overall Cronbach’s α coefficient reached 0.921, indicating excellent internal reliability. Subscale coefficients were also robust, with values of 0.894 for PANAS, 0.848 for SRF-S, and 0.935 for CDS, thereby demonstrating strong reliability across all measurement tools. Internal consistency reliability was classified based on conventional thresholds: coefficients above 0.8 were considered excellent, those between 0.6 and 0.8 acceptable, and those below 0.6 inadequate27. In the formal survey phase, all investigators received standardized training prior to data collection. Data collection was conducted by trained investigators following the same standardized protocol, with a designated lead investigator supervising adherence to the procedures. Training focused on eligibility screening, standardized administration of the questionnaires, confidentiality, and accurate data entry according to the study protocol. During formal data collection, investigators used the same outpatient office and adhered strictly to standardized protocols. General information was collected to describe participants’ health and sociodemographic profiles and to support exploratory subgroup comparisons. These variables included age, gender, smoking history, drinking history, body mass index (BMI), cataract nuclear (N), cortical (C), and posterior subcapsular (P) grading, disease duration, number and laterality of affected eyes, visual acuity of the affected eye, education level, monthly income per capita, marital status, living alone, and family history of cataract. BMI was calculated as weight (kg) divided by height squared (m2) based on measured height and weight. Data collection was conducted using a mixed-mode approach. Sex, age, and all scale items (PANAS, SRF-S, and CDS) were completed directly using the Wenjuanxing online platform, which required mandatory completion before submission. Sociodemographic and clinical variables were collected using a paper-based general information form during face-to-face interviews and were subsequently manually entered into the Wenjuanxing system by trained investigators. The remaining data were collected through face-to-face interviews, during which investigators recorded responses to patients’ answers. All scale items (PANAS, SRF-S, and CDS) were self-reported by participants; investigators read each item aloud when needed and recorded the participant’s selected response to ensure comprehension, given older age and visual impairment. Standardized instructions were provided to all participants explaining the study’s purpose and procedures. Data were recorded in real time using the Wenjuanxing online platform and stored securely in digital format. Quality assurance measures To ensure methodological rigor, all measurement instruments were selected based on established psychometric validity and confirmed to have high internal reliability within the target population. A standardized, interviewer-administered protocol was used to consistently collect questionnaire responses. Each face-to-face interview was conducted by one investigator specifically trained to be impartial, especially when posing delicate questions concerning “care dependence” or emotional discomfort. The interviewer recorded the participant’s self-reported answers in real time. Questionnaires with excessive missing values, patterned responses, or logical inconsistencies were removed. Two independent statisticians performed data verification and cleaning, with missing values addressed using multiple imputation. The database was subsequently locked after all data discrepancies were resolved, thereby ensuring the completeness and authenticity of the final dataset. Sample size calculation In accordance with Kendall’s principle for multivariate analysis, the recommended sample size should be five to ten times the number of observed variables included in the questionnaire28. To enhance the robustness of the study and accommodate potential attrition or data loss due to unforeseen circumstances, an additional 10–20% contingency margin was incorporated. The general information questionnaire included 18 variables, while PANAS, SRF-S, and CDS comprised 18, 16, and 15 items, respectively. Based on a conservative estimate of 5 times the total number of items and a 10% adjustment for non-response or invalid data, the minimum required sample size was approximately 369 participants. Kline29 recommends an N:q ratio of approximately 20:1 (participants per estimated parameter), with 10:1 as a less favorable lower bound, implying that models with 20–40 free parameters typically require samples of 200–400 or more, particularly when maximum likelihood estimation is used. Wolf et al.30 used Monte Carlo simulations and showed that required sample sizes for common structural equation modeling (SEM) can range from about 30 to 460 cases, depending on the number of indicators, factor loadings, and model complexity, with larger samples needed when loadings are modest and the model is more complex. Given the model structure and factor loadings, a sample of 452 lies at the upper end of this empirically derived range and thus provides sufficient power and stable parameter estimates for the SEM. Statistical analysis Data management, including linking the clinical ophthalmic dataset and the questionnaire dataset by patient name, verifying duplicates by patient name, and performing data cleaning and questionnaire scoring, was performed in R version 4.4.3 (The R Foundation for Statistical Computing, Vienna, Austria). Statistical analyses were conducted using SPSS version 27.0 (IBM Corp., Armonk, NY, USA) and AMOS version 26.0 (IBM Corp., Armonk, NY, USA). Normality testing was applied to the total scores of positive affect, negative affect, self-regulatory fatigue, and care dependency. Because these scale total scores were not normally distributed, they were summarized as medians with interquartile ranges (IQR) and compared using non-parametric tests. The Mann-Whitney U test was applied for comparisons between two groups, while the Kruskal-Wallis test was used for comparisons involving three or more groups. Categorical variables were presented as frequencies and percentages (n [%]). Group comparisons for categorical variables were evaluated using the chi-square test or Fisher’s exact test as appropriate. Spearman’s rank correlation coefficients were calculated to examine the associations among positive affect, negative affect, self-regulatory fatigue, and care dependency. Path analysis was conducted using SEM in AMOS to assess the mediating role of self-regulatory fatigue in the relationship between emotional states and care dependency. For the primary SEM, a parsimonious, theory‑driven model was specified, including positive affect and negative affect as exogenous variables, self‑regulatory fatigue as the mediator, and care dependency (CDS total score) as the outcome. Age and best‑corrected visual acuity of the affected eye were included as covariates based on their established clinical relevance to functional status. Other demographic variables (e.g., gender, marital status, income) were not included as direct predictors in the primary model to minimize model complexity and reduce the risk of overfitting. An extended model including additional demographic covariates was examined as a sensitivity analysis and is presented in the Supplementary Materials. The models are grounded in stress-coping and affect-regulation frameworks, in which self‑regulating fatigue, emotional states, and functional limitations mutually influence adaptation to chronic visual impairment and related care needs. Accordingly, care dependency was modeled as the primary outcome, directly influenced by age, gender, marital status, visual acuity of the affected eye, and self‑regulating fatigue, reflecting well‑established evidence that older age, female gender, being unmarried, and poorer vision are associated with greater disability and need for assistance, and that higher fatigue undermines functional autonomy. Positive and negative affect were specified as both predictors and potential mediators because prior literature shows that socioeconomic resources (e.g., monthly income per capita) and clinical factors shape emotional well‑being, which in turn affects perceived disability, coping, and engagement in self‑care. Visual acuity, age, and gender were allowed to influence self‑regulating fatigue and affective states, based on models of chronic illness adjustment, indicating that greater disease severity and sociodemographic vulnerability are linked to higher symptom burden and more negative emotional responses. Finally, correlations among exogenous variables and residuals were specified where theory and prior empirical work suggest shared but unmeasured determinants (e.g., socioeconomic and social support factors jointly affecting income, marital status, and gender-role-related patterns of care). Model fit was assessed using CMIN/DF (chi-square/degree of freedom), RFI (Relative Fit Index), IFI (Incremental Fit Index), TLI (Tucker-Lewis Index), CFI (Comparative Fit Index), and RMSEA (Root Mean Square Error of Approximation). Because the total scores for positive affect, negative affect, self‑regulatory fatigue, and care dependency deviated from normality, bootstrapping procedures were used in AMOS to obtain robust estimates for the SEM. Specifically, 1000 bootstrap samples were generated, and 95% bias‑corrected bootstrap confidence intervals were calculated for all structural path coefficients. This procedure provides robust parameter estimates and significance tests that do not rely on the assumption of multivariate normality and are consistent with recommended practice for SEM with non‑normal data. Bias-corrected 95% confidence intervals (CI) were generated to evaluate mediation effects. All statistical tests were two-tailed, and a P-value of less than 0.05 was considered statistically significant. Results Initially, this study collected 512 questionnaires. The clinical ophthalmic dataset containing eye-related clinical information and the self-completed questionnaire dataset were linked by matching the patient’s name across the two datasets to integrate clinical eye information with questionnaire responses; this yielded 507 linked records for analysis. The following cases were excluded: (1) 6 duplicate names identified during verification of the linked dataset (only one record retained per duplicate); (2) 1 case with missing data; (3) 30 cases with abnormal CNP grading (N4 and N5) in both eyes; (4) 15 cases with visual acuity above 0.5 in both eyes; (5) 1 patient unwilling to disclose household income; (6) 1 patient uncertain about whether they were living alone; (7) 1 patient who selected both “none” and “heart disease” under comorbidities (logical error). A total of 452 valid questionnaires were finally included. Demographic characteristics Among the 452 participants, the median age was 72 years, and most were female. The median (IQR) scores were 27 (20, 28) for positive affect, 25 (20, 28) for negative affect, 47.5 (42, 53) for self-regulatory fatigue, and 64 (57, 68) for care dependency (Table 1). Table 1 shows the variables with statistically significant differences in positive affect, negative affect, self-regulatory fatigue, and CDS. BMI, living alone, disease duration, and cataract CNP grading were not associated with differences in positive affect, negative affect, self-regulatory fatigue, or CDS (all P > 0.05), except for a general trend toward lower positive affect and CDS with increasing N grade (P = 0.013 and < 0.001, respectively). BMI was analyzed as a general indicator of nutritional status and overall physical health because both underweight and obesity in older adults have been linked to functional limitation and disability in activities of daily living, which may relate to care dependence. BMI has also been associated with psychological distress and fatigue-related conditions; therefore, it was collected to characterize the sample and to provide a basis for considering potential confounding in future analyses. The two multiple-response baseline items are summarized in Fig. 1. Table 1Demographic characteristics.VariablesN (%)Positive affectNegative affectSelf-regulatory fatigueCare dependencyMedian (P25, P75) P Median (P25, P75) P Median (P25, P75) P Median (P25,P75) P N = 45227 (20, 28)25 (20, 28)47.5 (42, 53)64 (57, 68)Age72 (67, 79) Gender 0.031< 0.001< 0.0010.266Male158 (34.96)27 (20, 29)23 (18, 27)45 (39, 51)64 (57, 69)Female294 (65.04)27 (20, 28)26 (21, 29)48 (43, 54)63 (57, 68) Education level 0.002< 0.001< 0.001< 0.001Illiterate31 (6.86)24 (19, 27)27 (24, 31)54 (45, 57)57 (52, 64)Primary school85 (18.81)25 (20, 28)26 (21.5, 28)48 (44, 53)62 (55, 67)Junior high school105 (23.23)27 (20, 28)26 (21, 28)48 (44.5, 53)61 (56, 67)Technical secondary school79 (17.48)27 (21, 29)25 (20, 28)48 (40, 54)65 (60, 69)High school63 (13.94)27 (23, 29)23 (19, 27)46 (42, 50)66 (61, 69)Associate degree37 (8.19)28 (25, 30.5)20 (17.5, 24)42 (37, 47.5)67 (62, 70)Bachelor’s degree or above52 (11.5)27 (20, 30)23 (18, 29)45 (39, 51)67 (57, 71) Marital status < 0.001< 0.001< 0.001< 0.001Married333 (73.67)27 (21, 29)24 (19, 28)46 (41, 52)65 (59, 69)Other119 (26.33)24 (19, 27)27 (22, 30)50 (45, 55)60 (53, 66) Monthly income per capita < 0.0010.004< 0.001< 0.001< 200097 (21.46)24 (19, 27)27 (21.5, 30)51 (44, 56)60 (54, 65)2000–4999278 (61.5)27 (21, 28)24 (20, 28)48 (42, 52)64 (59, 69)>=500077 (17.04)27 (21, 29.5)23 (18, 28)45 (39, 50.5)66 (57, 72) Smoking history 0.0300.0080.0060.228Yes117 (25.88)27 (20, 29)23 (18, 27)46 (40, 52)64 (57, 69.5)No335 (74.12)27 (20, 28)25 (21, 28)48 (43, 54)63 (57, 68) Drinking history 0.2980.0350.0470.999Yes124 (27.43)27 (20, 29)23 (19v27.75)46 (40, 52)64 (56.25, 69)No328 (72.57)27 (21, 28)25 (20, 28)48 (43, 53)63.5 (58, 68) Family history of cataract < 0.0010.019< 0.001< 0.001Yes83 (18.36)28 (24, 30)23 (18, 27)45 (40, 49)67 (62, 70)No106 (23.45)27 (22, 29)24 (19, 28)45.5 (39.75, 52)66 (59, 69.25)Unknown263 (58.19)26 (20, 28)25 (21, 28)49 (43, 54)62 (55, 67) Number and laterality of affected eyes 0.0780.0010.0050.029Unilateral36 (7.96)27 (24.25, 29)21 (17, 25)42.5 (39, 50)67 (59.75, 70)Bilateral416 (92.04)27 (20, 28)25 (20, 28)48 (42, 53)63 (57, 68) Visual acuity of affected eye < 0.001< 0.001< 0.001< 0.001Mild109 (24.12)28 (27, 30)20 (16, 24)43 (39, 48)68 (63, 70.5)Moderate166 (36.73)27 (21, 28)25 (20, 28)47 (42, 52)64 (59, 68)Severe52 (11.5)27 (20, 28)27 (23, 28)49 (44.25, 52.75)62 (57.25, 67)Blind125 (27.65)21 (19, 27)27 (22.5, 30)52 (46, 56)57 (51, 66)The highest-income category was merged with the preceding group due to the small size of the subgroup.Variables were presented in a structured order (demographic, socioeconomic/lifestyle, disease-related, and clinical severity) to align with conventional reporting practice and improve interpretability in observational studies. Table 1Demographic characteristics.VariablesN (%)Positive affectNegative affectSelf-regulatory fatigueCare dependencyMedian (P25, P75) P Median (P25, P75) P Median (P25, P75) P Median (P25,P75) P N = 45227 (20, 28)25 (20, 28)47.5 (42, 53)64 (57, 68)Age72 (67, 79) Gender 0.031< 0.001< 0.0010.266Male158 (34.96)27 (20, 29)23 (18, 27)45 (39, 51)64 (57, 69)Female294 (65.04)27 (20, 28)26 (21, 29)48 (43, 54)63 (57, 68) Education level 0.002< 0.001< 0.001< 0.001Illiterate31 (6.86)24 (19, 27)27 (24, 31)54 (45, 57)57 (52, 64)Primary school85 (18.81)25 (20, 28)26 (21.5, 28)48 (44, 53)62 (55, 67)Junior high school105 (23.23)27 (20, 28)26 (21, 28)48 (44.5, 53)61 (56, 67)Technical secondary school79 (17.48)27 (21, 29)25 (20, 28)48 (40, 54)65 (60, 69)High school63 (13.94)27 (23, 29)23 (19, 27)46 (42, 50)66 (61, 69)Associate degree37 (8.19)28 (25, 30.5)20 (17.5, 24)42 (37, 47.5)67 (62, 70)Bachelor’s degree or above52 (11.5)27 (20, 30)23 (18, 29)45 (39, 51)67 (57, 71) Marital status < 0.001< 0.001< 0.001< 0.001Married333 (73.67)27 (21, 29)24 (19, 28)46 (41, 52)65 (59, 69)Other119 (26.33)24 (19, 27)27 (22, 30)50 (45, 55)60 (53, 66) Monthly income per capita < 0.0010.004< 0.001< 0.001< 200097 (21.46)24 (19, 27)27 (21.5, 30)51 (44, 56)60 (54, 65)2000–4999278 (61.5)27 (21, 28)24 (20, 28)48 (42, 52)64 (59, 69)>=500077 (17.04)27 (21, 29.5)23 (18, 28)45 (39, 50.5)66 (57, 72) Smoking history 0.0300.0080.0060.228Yes117 (25.88)27 (20, 29)23 (18, 27)46 (40, 52)64 (57, 69.5)No335 (74.12)27 (20, 28)25 (21, 28)48 (43, 54)63 (57, 68) Drinking history 0.2980.0350.0470.999Yes124 (27.43)27 (20, 29)23 (19v27.75)46 (40, 52)64 (56.25, 69)No328 (72.57)27 (21, 28)25 (20, 28)48 (43, 53)63.5 (58, 68) Family history of cataract < 0.0010.019< 0.001< 0.001Yes83 (18.36)28 (24, 30)23 (18, 27)45 (40, 49)67 (62, 70)No106 (23.45)27 (22, 29)24 (19, 28)45.5 (39.75, 52)66 (59, 69.25)Unknown263 (58.19)26 (20, 28)25 (21, 28)49 (43, 54)62 (55, 67) Number and laterality of affected eyes 0.0780.0010.0050.029Unilateral36 (7.96)27 (24.25, 29)21 (17, 25)42.5 (39, 50)67 (59.75, 70)Bilateral416 (92.04)27 (20, 28)25 (20, 28)48 (42, 53)63 (57, 68) Visual acuity of affected eye < 0.001< 0.001< 0.001< 0.001Mild109 (24.12)28 (27, 30)20 (16, 24)43 (39, 48)68 (63, 70.5)Moderate166 (36.73)27 (21, 28)25 (20, 28)47 (42, 52)64 (59, 68)Severe52 (11.5)27 (20, 28)27 (23, 28)49 (44.25, 52.75)62 (57.25, 67)Blind125 (27.65)21 (19, 27)27 (22.5, 30)52 (46, 56)57 (51, 66)The highest-income category was merged with the preceding group due to the small size of the subgroup.Variables were presented in a structured order (demographic, socioeconomic/lifestyle, disease-related, and clinical severity) to align with conventional reporting practice and improve interpretability in observational studies. Demographic characteristics. The highest-income category was merged with the preceding group due to the small size of the subgroup. Variables were presented in a structured order (demographic, socioeconomic/lifestyle, disease-related, and clinical severity) to align with conventional reporting practice and improve interpretability in observational studies. Fig. 1Distribution of comorbidities and medication history among elderly patients with senile cataract (multiple choices allowed). Distribution of comorbidities and medication history among elderly patients with senile cataract (multiple choices allowed). Correlation analysis All correlation analyses were conducted using the total questionnaire scores of PANAS, SRF-S, and CDS. Positive and negative affect were moderately inversely correlated (r = −0.611, P < 0.001). Positive affect showed strong negative associations with self-regulatory fatigue (r = −0.738, P < 0.001) and a strong positive association with care dependency (r = 0.697, P < 0.001). Negative affect was positively correlated with self-regulatory fatigue (r = 0.783, P < 0.001) and negatively correlated with care dependency (r = −0.602, P < 0.001). Self-regulatory fatigue was negatively correlated with care dependency (r = −0.705, P < 0.001) (Table 2). These Spearman correlations are unadjusted bivariate associations; the subsequent SEM path analysis estimated the associations while accounting for the visual acuity of the affected eye. For the CDS, higher total scores indicate greater independence and lower care dependence. Table 2Correlation analysis.Positive affectNegative affectSelf-regulatory fatigueCare dependencyPositive affect1Negative affect− 0.611 (P < 0.001)1Self-regulatory fatigue− 0.738 (P < 0.001)0.783 (P < 0.001)1Care dependency0.697 (P < 0.001)− 0.602 (P < 0.001)− 0.705 (P < 0.001)1Higher CDS score indicates greater independence (lower care dependency). Table 2Correlation analysis.Positive affectNegative affectSelf-regulatory fatigueCare dependencyPositive affect1Negative affect− 0.611 (P < 0.001)1Self-regulatory fatigue− 0.738 (P < 0.001)0.783 (P < 0.001)1Care dependency0.697 (P < 0.001)− 0.602 (P < 0.001)− 0.705 (P < 0.001)1Higher CDS score indicates greater independence (lower care dependency). Correlation analysis. Higher CDS score indicates greater independence (lower care dependency). Path analysis The parsimonious SEM path analysis was performed using the total scores from the PANAS, SRF-S, and CDS questionnaires (Fig. 2). Table 3 shows that the parsimonious model had a good fit. In the parsimonious model, positive affect was positively associated with the CDS score (standardized β = 0.324, P < 0.001; higher CDS scores indicate greater independence/lower care dependence) and negatively associated with self-regulatory fatigue (standardized β = − 0.367, P < 0.001). In contrast, negative affect was negatively associated with the CDS score (standardized β = − 0.139, P = 0.004), indicating greater care dependence, and positively associated with self-regulatory fatigue (standardized β = 0.543, P < 0.001). Self-regulatory fatigue was negatively associated with the CDS score (standardized β = − 0.184, P = 0.001), also indicating greater care dependence (Table 4). Bootstrap mediation analysis confirmed significant indirect effects on the CDS score through self-regulatory fatigue for positive affect (standardized β = 0.067, 95% CI 0.027–0.109, P = 0.002) and negative affect (standardized β = − 0.100, 95% CI − 0.156 to − 0.041, P = 0.003) (Table 5). The results reported here are based on the parsimonious SEM, which includes only age and visual acuity as covariates; results from the extended model with additional demographic covariates showed a similar pattern of associations, but that more complex model must be considered exploratory. Fig. 2Parsimonious mediation model. Higher CDS scores indicate greater independence. Parsimonious mediation model. Higher CDS scores indicate greater independence. Table 3Model fit of the parsimonious path analysis.ReferenceActualCMIN/DF< 51.853RFI> 0.90.979IFI> 0.90.999TLI> 0.90.990CFI> 0.90.999RMSEA< 0.080.043CMIN/DF: chi-square value divided by its degrees of freedom; RFI: relative fit index; IFI: incremental fit index; TLI: Tucker-Lewis index; CFI: comparative fit index; RMSEA: root mean square error of approximation. Table 3Model fit of the parsimonious path analysis.ReferenceActualCMIN/DF< 51.853RFI> 0.90.979IFI> 0.90.999TLI> 0.90.990CFI> 0.90.999RMSEA< 0.080.043CMIN/DF: chi-square value divided by its degrees of freedom; RFI: relative fit index; IFI: incremental fit index; TLI: Tucker-Lewis index; CFI: comparative fit index; RMSEA: root mean square error of approximation. Model fit of the parsimonious path analysis. CMIN/DF: chi-square value divided by its degrees of freedom; RFI: relative fit index; IFI: incremental fit index; TLI: Tucker-Lewis index; CFI: comparative fit index; RMSEA: root mean square error of approximation. Table 4Mediation model: Path coefficients of the parsimonious SEM model.EstimateStandardized EstimateS.E.C.R. P Positive affect<---Visual acuity of affected eye− 1.486− 0.3490.177− 8.381< 0.001Negative affect<---Visual acuity of affected eye1.6410.3190.2247.333< 0.001Positive affect<---Age− 0.183− 0.3090.025− 7.414< 0.001Negative affect<---Age0.1540.2140.0314.931< 0.001Self-regulating fatigue<---Negative affect0.6760.5430.03817.940< 0.001Self-regulating fatigue<---Positive affect− 0.552− 0.3670.047− 11.802< 0.001Self-regulating fatigue<---Age0.1020.1140.0234.393< 0.001Care dependency<---Positive affect0.5470.3240.0757.257< 0.001Care dependency<---Negative affect− 0.194− 0.1390.068− 2.8410.004Care dependency<---Visual acuity of affected eye− 0.945− 0.1310.237− 3.979< 0.001Care dependency<---Self-regulating fatigue− 0.206− 0.1840.065− 3.1820.001Care dependency<---Age− 0.271− 0.2710.033− 8.251< 0.001 Table 4Mediation model: Path coefficients of the parsimonious SEM model.EstimateStandardized EstimateS.E.C.R. P Positive affect<---Visual acuity of affected eye− 1.486− 0.3490.177− 8.381< 0.001Negative affect<---Visual acuity of affected eye1.6410.3190.2247.333< 0.001Positive affect<---Age− 0.183− 0.3090.025− 7.414< 0.001Negative affect<---Age0.1540.2140.0314.931< 0.001Self-regulating fatigue<---Negative affect0.6760.5430.03817.940< 0.001Self-regulating fatigue<---Positive affect− 0.552− 0.3670.047− 11.802< 0.001Self-regulating fatigue<---Age0.1020.1140.0234.393< 0.001Care dependency<---Positive affect0.5470.3240.0757.257< 0.001Care dependency<---Negative affect− 0.194− 0.1390.068− 2.8410.004Care dependency<---Visual acuity of affected eye− 0.945− 0.1310.237− 3.979< 0.001Care dependency<---Self-regulating fatigue− 0.206− 0.1840.065− 3.1820.001Care dependency<---Age− 0.271− 0.2710.033− 8.251< 0.001 Mediation model: Path coefficients of the parsimonious SEM model. Table 5Bootstrap mediation analysis of the parsimonious SEM model.Model pathsStandardized Total effectsStandardized direct effectsStandardized indirect effectsβ (95% CI) P β (95% CI) P β (95% CI) P Age→Negative affect0.214 (0.129, 0.298)0.0020.214 (0.129,0.298)0.002Visual acuity of affected eye→Negative affect0.319 (0.227, 0.404)0.0030.319 (0.227, 0.404)0.003Age→Positive affect− 0.309 (− 0.392, − 0.223)0.003− 0.309 (− 0.392, − 0.223)0.003Visual acuity of affected eye →Positive affect− 0.349 (− 0.426, − 0.261)0.003− 0.349 (− 0.426, − 0.261)0.003Age→Self-regulating fatigue0.344 (0.262, 0.423)0.0020.114 (0.060, 0.161)0.0040.230 (0.163, 0.295)0.002Visual acuity of affected eye→Self-regulating fatigue0.301 (0.225, 0.369)0.0020.301 (0.225, 0.369)0.002Negative affect→Self-regulating fatigue0.543 (0.471, 0.604)0.0020.543 (0.471, 0.604)0.002Positive affect→Self-regulating fatigue− 0.367 (− 0.435, − 0.297)0.002− 0.367 (− 0.435, − 0.297)0.002Age→Care dependency− 0.464 (− 0.543, − 0.387)0.002− 0.271 (− 0.348, − 0.199)0.002− 0.193 (− 0.250, − 0.143)0.002Visual acuity of affected eye→Care dependency− 0.344 (− 0.416, − 0.266)0.003− 0.131 (− 0.194, − 0.068)0.003− 0.213 (− 0.267, − 0.163)0.002Negative affect→Care dependency− 0.239 (− 0.321, − 0.143)0.003− 0.139 (− 0.224, − 0.049)0.003− 0.100 (− 0.156, − 0.041)0.003Positive affect→Care dependency0.391 (0.312, 0.470)0.0010.324 (0.236, 0.417)0.0010.067 (0.027, 0.109)0.002Self-regulating fatigue→Care dependency− 0.184 (− 0.283, − 0.075)0.003− 0.184 (− 0.283, − 0.075)0.003 Table 5Bootstrap mediation analysis of the parsimonious SEM model.Model pathsStandardized Total effectsStandardized direct effectsStandardized indirect effectsβ (95% CI) P β (95% CI) P β (95% CI) P Age→Negative affect0.214 (0.129, 0.298)0.0020.214 (0.129,0.298)0.002Visual acuity of affected eye→Negative affect0.319 (0.227, 0.404)0.0030.319 (0.227, 0.404)0.003Age→Positive affect− 0.309 (− 0.392, − 0.223)0.003− 0.309 (− 0.392, − 0.223)0.003Visual acuity of affected eye →Positive affect− 0.349 (− 0.426, − 0.261)0.003− 0.349 (− 0.426, − 0.261)0.003Age→Self-regulating fatigue0.344 (0.262, 0.423)0.0020.114 (0.060, 0.161)0.0040.230 (0.163, 0.295)0.002Visual acuity of affected eye→Self-regulating fatigue0.301 (0.225, 0.369)0.0020.301 (0.225, 0.369)0.002Negative affect→Self-regulating fatigue0.543 (0.471, 0.604)0.0020.543 (0.471, 0.604)0.002Positive affect→Self-regulating fatigue− 0.367 (− 0.435, − 0.297)0.002− 0.367 (− 0.435, − 0.297)0.002Age→Care dependency− 0.464 (− 0.543, − 0.387)0.002− 0.271 (− 0.348, − 0.199)0.002− 0.193 (− 0.250, − 0.143)0.002Visual acuity of affected eye→Care dependency− 0.344 (− 0.416, − 0.266)0.003− 0.131 (− 0.194, − 0.068)0.003− 0.213 (− 0.267, − 0.163)0.002Negative affect→Care dependency− 0.239 (− 0.321, − 0.143)0.003− 0.139 (− 0.224, − 0.049)0.003− 0.100 (− 0.156, − 0.041)0.003Positive affect→Care dependency0.391 (0.312, 0.470)0.0010.324 (0.236, 0.417)0.0010.067 (0.027, 0.109)0.002Self-regulating fatigue→Care dependency− 0.184 (− 0.283, − 0.075)0.003− 0.184 (− 0.283, − 0.075)0.003 Bootstrap mediation analysis of the parsimonious SEM model. In the complex model (Supplementary Fig. S1), model fit indices were CMIN/DF = 3.073, RFI = 0.925, IFI = 0.977, TLI = 0.948, CFI = 0.977, and RMSEA = 0.068 (Supplementary Table S1). Positive affect was positively associated with the CDS score (standardized β = 0.319, P < 0.001; higher CDS scores indicate greater independence/lower care dependence) and negatively associated with self-regulatory fatigue (standardized β = − 0.360, P < 0.001). In contrast, negative affect was negatively associated with the CDS score (standardized β = − 0.129, P = 0.008), indicating greater care dependence, and positively associated with self-regulatory fatigue (standardized β = 0.521, P < 0.001). Self-regulatory fatigue was negatively associated with the CDS score (standardized β = − 0.190, P = 0.001), also indicating greater care dependence (Supplementary Table S2). Bootstrap mediation analysis confirmed significant indirect effects on the CDS score through self-regulatory fatigue for positive affect (standardized β = 0.068, 95% CI 0.028–0.109, P = 0.003) and negative affect (standardized β = − 0.099, 95% CI − 0.152 to − 0.043, P = 0.002) (Supplementary Table S3). Because both the direct paths from positive affect to the CDS score and from negative affect to the CDS score remained significant after including self-regulatory fatigue, the results indicate partial mediation in the emotion-to-self-regulatory fatigue-to-care dependence pathway (Tables 3 and 4). The overall model demonstrated acceptable fit (Supplementary Table S1). Discussion This study examined the associations among positive affect, negative affect, self-regulatory fatigue, and care dependence in elderly patients with senile cataract, using structural equation modeling to test a theoretically grounded mediation pathway. The findings revealed that self-regulatory fatigue partially mediated the relationships between both positive and negative affect and care dependence, while significant direct effects of affect on care dependence persisted. These results extend prior work in geriatric populations, where emotional resilience has been associated with reduced risk of functional decline in older adults facing sensory impairment, and advance understanding of the psychological mechanisms through which affective states shape functional outcomes in the specific context of age-related visual impairment. Positive affect pathway The finding that positive affect was associated with greater care independence, with self-regulatory fatigue as a partial mediator, is consistent with and extends the growing literature linking positive affect to functional outcomes in older adults. Fox et al31. demonstrated in 277 community-dwelling older adults with subjective cognitive decline that positive affect was independently associated with better functional ability even after controlling for objective cognitive performance, suggesting that positive affect may exert influence over and above cognitive resources. The present study extends this observation to elderly cataract patients, a population that faces specific compensatory cognitive burdens imposed by visual impairment, and further identifies self-regulatory fatigue as a mechanism through which this association partly operates. From a theoretical standpoint, the broaden-and-build model proposes that positive affect expands the cognitive-behavioral repertoire, supporting more flexible and adaptive coping responses and thus conserving regulatory resources7,8. The partial mediation pattern observed here, where a substantial direct effect of positive affect on CDS scores persisted after including SRF, is consistent with this model’s prediction of multiple parallel pathways. Beyond reduced self-regulatory fatigue, positive affect may support care independence through enhanced perceived social support, greater activity engagement, and stronger coping self-efficacy7,11. These alternative pathways were not directly measured in the present study but warrant investigation in future research. Negative affect pathway The association between higher negative affect and greater care dependence, partly mediated by elevated self-regulatory fatigue, converges with evidence from studies of comparable populations. The systematic review and meta-analysis by Goodarzi et al32. identified psychosocial factors, particularly depressive affect, as the strongest determinants of functional dependence among older adults, with depression associated with a 76% increase in functional dependence across 46 studies. In the more focused context of elderly nursing facility applicants in China, Ding et al33. found that each 1-point decrease in ADL score was associated with a 6% increase in depression prevalence, illustrating the bidirectional and reinforcing nature of the affect-functional dependence relationship in comparable East Asian populations. The present finding that negative affect predicted lower CDS scores (greater dependence) both directly and via elevated SRF provides mechanistic specificity to these associations: rather than a general mood effect, the pathway from negative affect to dependence is at least partly explained by the depletion of self-regulatory resources through sustained emotional monitoring and suppression demands13–15. Moreover, in the ophthalmic context specifically, Eramudugolla et al34. found that depression and anxiety in cataract and age-related macular degeneration patients were significantly associated with objective measures of visual function independently of self-reported eye disease. The present study complements this work by demonstrating that it is not merely the presence of negative affect but its downstream effect on regulatory capacity that further undermines functional independence. Because a significant direct effect of negative affect on care dependence remained after accounting for SRF, other mechanisms, such as activity avoidance, reduced motivation to initiate self-care routines, and lower perceived social support, are also likely to contribute, consistent with Thoits’ model of the health consequences of social stress11. Role of self-regulatory fatigue Self-regulatory fatigue emerged as a robust partial mediator in both pathways, underscoring its central importance in the affect-care dependence relationship. This finding situates the present work within a growing body of evidence demonstrating the clinical relevance of the SRF construct across a range of chronic conditions. In a sample of 338 older adults with chronic heart failure, Li et al.35 showed that social support influenced self-regulatory fatigue through symptom burden, psychological resilience, and rumination, collectively accounting for 52.2% of the total effect. Similarly, Wu et al.36 reported that self-regulatory fatigue partially mediated the relationship between perceived stress and self-management behaviors in adults living with HIV, explaining 34.4% of the total effect. The present findings are consistent with these results and extend the relevance of SRF-mediated pathways to care dependence outcomes, specifically, in a population where compensatory cognitive load from visual impairment may uniquely accelerate regulatory resource depletion16,17. The magnitude of the SRF→CDS path coefficient was somewhat smaller than the direct affect paths, indicating that while SRF is a meaningful mechanism, it accounts for a modest proportion of the total effect. This is in line with Wang et al.’s37 demonstration in type 2 diabetes patients that self-regulatory resource depletion was associated with poorer self-care in exercise and diet domains, but that the relationship was context- and domain-specific rather than globally determining. In the cataract context, the additional burden of compensatory visual processing may create a distinct regulatory demand that future studies should assess with more granular measures of cognitive workload. Notably, Solberg Nes et al.38 found in hematologic malignancy patients undergoing transplantation that pre-transplant self-regulatory fatigue predicted both quality of life and health behavior outcomes one year later, supporting the longitudinal prognostic relevance of SRF in medically vulnerable populations. Replication of the present cross-sectional findings with a longitudinal design would be a valuable next step. Clinical implications The identification of self-regulatory fatigue as a modifiable partial mediator, together with the independent direct effects of both positive and negative affect, points to multiple intervention leverage points. Interventions targeting positive affect enhancement (e.g., mindfulness-based stress reduction, brief gratitude journaling, verbal sharing groups) are consistent with evidence from meta-analyses demonstrating that positive psychology interventions significantly improve well-being and reduce depressive symptoms in community samples. For the SRF pathway specifically, tailored energy management programs and self-regulation enhancement strategies have demonstrated feasibility in chronic illness populations18,39, and their adaptation for elderly cataract patients, with particular attention to visual accessibility and literacy levels, merits evaluation in future controlled studies. The finding that direct affect effects persist even after accounting for SRF additionally suggests that approaches targeting emotional states directly, such as pre-surgical psychological counseling, may independently support functional outcomes beyond their effect on fatigue. From a care delivery standpoint, the present findings support the routine assessment of affective states and self-regulatory fatigue in ophthalmology outpatient settings. Such assessment could facilitate earlier identification of patients at elevated risk of care dependence prior to (or independent of) surgical intervention, complementing existing recommendations for mental health screening in visually impaired older adults40. Limitations Several limitations warrant consideration. First, the cross-sectional design precludes causal inference; the pathways identified should therefore be interpreted as plausible mechanisms rather than directional effects. Second, the sample was drawn from a single tertiary hospital in Southwest China, which may limit generalizability to populations with differing sociocultural or healthcare contexts. Third, although visual acuity was adjusted for in the SEM, residual confounding by physical impairment cannot be excluded, and functional restrictions related to reduced visual acuity may account for a larger proportion of care dependence than affective status alone. Fourth, data were collected via face-to-face, investigator-administered interviews, which may have introduced social desirability bias; standardized training and neutral interviewing procedures were implemented to minimize this risk, but bias cannot be fully excluded. . Fifth, cognitive impairment was screened using clinical judgment based on medical records and interview rather than formal instruments such as the Mini-Mental State Examination or the Montreal Cognitive Assessment, which may have introduced classification error. Finally, sample size was determined using a rule-of-thumb approach rather than an SEM-specific power analysis; although the final sample (n = 452) was relatively large, future studies should consider power-based sample size determination tailored to the planned path model. Clinically meaningful thresholds for PANAS, SRF-S, and CDS score changes in this population are not yet well established, and the clinical significance of small score differences should be interpreted cautiously until future studies link score changes to external clinical outcomes. Conclusion In elderly patients with senile cataract, positive affect was associated with greater care independence and negative affect with greater care dependence, with self-regulatory fatigue serving as a partial mediator in both pathways. These findings extend and contextualize prior work on psychosocial determinants of functional independence in older adults and in chronic illness populations where self-regulatory fatigue has been implicated as a mechanism linking psychological states to health behaviors and outcomes. Routine assessment of affect and self-regulatory fatigue may help identify patients at risk of greater care dependence, and targeted interventions addressing these psychological factors warrant further evaluation in longitudinal and interventional designs. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1