The current study analyzed the latest dataset from a large and ongoing project conducted in southwest China, which focuses on physical and psychological well-being in cancer patients. A two-stage random sampling strategy with probability proportionate to sample size (PPS) design was performed to produce participants. In the first stage, the third affiliated hospital of Kunming Medical University was randomly selected from all hospitals with cancer patients; in the second stage, based on the estimated sample size, a total of 9 inpatient departments were randomly selected, and all initially included patients were selected from these units. Following the sampling results, a cross-sectional study was conducted from September to December 2023 in the third affiliated hospital of Kunming Medical University, which is one of the largest cancer centers in southwest China.

Eligibility criteria included being aged 18 or older, having a confirmed pathological diagnosis of primary cancer, and being able to read and understand the questionnaire items. Patients were excluded if they had (1) severe mental health disorders, (2) any physical illness preventing participation in the investigation, or (3) observed communication problems.

A self-designed questionnaire was adopted to collect clinical and sociodemographic features. The participants were asked about their age, gender, ethnicity, employment status, educational attainment, place of residence, marital status, financial burden, religious or spiritual beliefs, types of cancer, cancer stage, and coexisting medical conditions. Religious or spiritual beliefs were assessed using the question: “Do you have any religious or spiritual beliefs?” with response options of "Yes" or "No". Financial burden was evaluated with the question: “What is your household's financial burden?” with response options including "slight financial burden", "moderate financial burden", and "heavy financial burden". Participants were asked about coexisting medical conditions using the question: “Do you have any other medical conditions?” with response options of "Yes" or "No". In the current study, four variables (sex, ethnicity, educational attainment, and place of residence) were selected as social determinants.

The independent variables of the current study were sensory processing sensitivity and resilience. Sensory processing sensitivity was measured using the 10-item Chinese version of the Highly Sensitive Child Scale (CHSC) [19, 25]. Previous studies support the use of the CHSC for assessing adults, and its robustness and effectiveness have been validated [25]. The CHSC consists of ten items distributed across three dimensions: ease of excitation (EOE), aesthetic sensitivity (AES), and low sensory threshold (LST). This study is interested in the impact of negative environments. Therefore, the subscales EOE and LST, which assess sensitivity to negative environments, were used to measure the level of sensory processing sensitivity. Each item was rated on a scale ranging from 1 (not at all) to 7 (extremely). Higher scores indicate greater sensory processing sensitivity. The Cronbach’s α for CHSC, EOE, AES, and LST in the current study was 0.79 (95% CI: 0.77-0.81), 0.85 (95% CI: 0.83-0.87), 0.83 (95% CI: 0.80-0.86), and 0.86 (95% CI: 0.82-0.90), respectively. Psychological resilience was assessed using the Chinese version of the 10-item Connor-Davidson Resilience Scale (CD-RISC-10) [27]. The CD-RISC-10 is a validated and condensed version of the original 25-item CD-RISC. It consists of 10 items, each rated on a 5-point Likert scale ranging from 0 (never) to 4 (almost always). The combined score ranges from 0 to 40, with a higher score indicating better resilience. The Cronbach’s α for CD-RISC-10 in the current study was 0.91 (95% CI: 0.90-0.92). The total scores of CHSC and CD-RISC-10 were included in the analyses.

The primary outcome of the present study was the quality of life, which comprised four domains. The Chinese version of the World Health Organization Quality of Life-Brief (WHOQOL-BREF) was used to evaluate the quality of life of cancer patients [26]. This scale consists of 29 questions, with 24 items comprising four domains: physiology, psychology, social relationships, and environment. Except for one item, all questions were assessed on a 5-point Likert scale ranging from 1 (very poor) to 5 (very good). A high combined score indicates superior quality of life within the corresponding domains. The Cronbach’s α for the Chinese version of WHOQOL-BREF in the current study was 0.90 (95% CI: 0.89-0.91). The total scores of each domain of QoL were included in the analyses.

Preceding data acquisition, informed consent was received from all the participants. A self-administered questionnaire method was adopted for data collection. Participants were asked to complete the questionnaires independently, with adequate time and privacy. If participants needed assistance in completing the questionnaires, the investigators provided explanations. All investigators underwent pre-training and assessment. The clinical data, such as cancer types and staging, were further verified through reviewing medical records.

All data sorting and analysis were performed in the R statistical system (Version 4.3.2, The R Foundation for Statistical Computing, Vienna, Austria). The sociodemographic and clinical features of cancer patients were depicted using descriptive statistics. Categorical variables were presented as frequencies and percentages. Interval data characteristics were reported as mean with standard deviation (SD).

The associations between resilience, sensory processing sensitivity, and the four domains of quality of life were estimated using a linear regression model. Specifically, univariate linear regression was used to identify variables relevant to the four domains of quality of life. If the variables showed a significant correlation with the dependent variables (significance level set at less than 0.05, two-tailed), they were considered as covariates for subsequent analyses. In the second step, all the covariates and independent variables were included in multiple linear regression to estimate the associations between resilience, sensory processing sensitivity, and the four domains of quality of life (to address multiple testing, the significance level was adjusted to less than 0.0125, two-tailed). Subsequent sensitivity analyses were conducted using multiple linear regression with restricted cubic spline (RCS) to explore the changes in coefficients with increasing levels of resilience and sensory processing sensitivity. We hypothesize a linear dose-response relationship, where increasing resilience and sensory processing sensitivity linearly affect patients' quality of life, either improving or declining it.

Path analysis was adopted to examine whether resilience is a significant mediator between sensory processing sensitivity and quality of life. Specifically, we simultaneously entered sensory processing sensitivity, resilience, and the four domains of quality of life into the path model to estimate the standardized path coefficients (significance level set at less than 0.05, two-tailed). It is possible that the mediation effect of resilience may vary across different social characteristics. Consequently, we conducted subgroup path analyses to investigate whether social determinants such as sex, ethnicity, educational attainment, and place of residence moderate the mediation effect of resilience.

Ethical approval was obtained from the Institutional Review Committee of Yunnan Cancer Hospital (ky2021133). All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.