Psychometric assessment of the Chinese version of the perceptions of palliative care instrument in advanced cancer patients: a cross-sectional study

Human studies

The present study was established according to the ethical guidelines of the Helsinki Declaration and was approved by the Institutional Review Boards and Ethics Committee of Xinxiang Medical University (XXLL-2020B008).

Study design

This cross-sectional study validated the reliability and validity of the Chinese version of the perceptions of palliative care instrument (PPCI) using Brislin’s translation guidelines and established cross-cultural adaptation principles. Convenience sampling was employed for participant recruitment, with trained research assistants approaching potential participants in the inpatient wards and outpatient clinics of oncology and related departments. The assistants provided a detailed explanation of the study’s purpose and procedures, screened individuals against inclusion and exclusion criteria, and obtained written informed consent from those who agreed to participate. Data were collected via self-administered questionnaires, with assistance from trained investigators as needed.

Translation and cultural adaptation procedures

The PPCI was translated and culturally adapted into Chinese following Brislin’s guidelines and the cross-cultural adaptation process recommended by the American Academy of Orthopedic Surgeons Evidence-Based Medicine Committee (Jones et al., 2001; Beaton et al., 2000). Specifically, permission to develop the Chinese version of the PPCI (C-PPCI) was first obtained from the original authors and the authors received permission to use this instrument from the copyright holders. Forward translation was performed by two bilingual translators unfamiliar with the PPCI’s concepts and one researcher knowledgeable in palliative care terminology; the preliminary C-PPCI version was synthesized after discussions among the three translators. Back-translation was then conducted by two additional bilingual translators. Cultural adaptation was carried out by a panel of experts, including three palliative care specialists, one linguist, and one psychologist, who convened via online conference to evaluate the cultural appropriateness and accuracy of each item. The experts also assessed content validity through a two-round inquiry during this phase. Finally, a pilot test of the C-PPCI was conducted with 30 participants to evaluate its clarity and feasibility.

Settings and participants

A total of 537 patients with advanced cancer were recruited via convenience sampling from Xinxiang Central Hospital in Xinxiang, China, between October 2020 and March 2021. Inclusion criteria were as follows: (1) confirmed diagnosis of advanced-stage cancer, regardless of type; (2) aged 18 years or older; (3) able to understand the study’s purpose; and (4) willing to participate and provide written informed consent. Exclusion criteria included: (1) unconsciousness or disabilities involving literacy, hearing impairment, or inability to communicate; and (2) participation in other interventional studies. The sample size was calculated to be at least 444 participants, accounting for a requirement of 5–10 times the 37 items in the PPCI and an anticipated 20% attrition rate due to missing data. After explaining the study’s objectives, investigators obtained written informed consent from all participants. Data were collected through one-on-one administration, during which investigators distributed the questionnaires and offered assistance as needed; each completed questionnaire was immediately reviewed for completeness to minimize missing data.

The main measurement instruments

Statistical analysis

Data were analyzed using IBM SPSS version 22.0 and Mplus version 7.0, with the statistical significance level set at 0.05. Reliability was assessed via Cronbach’s alpha for internal consistency and test-retest reliability. Item analysis encompassed item discrimination, item-total correlations (≥0.30) (Streiner, Norman & Health, 2015), and critical ratio values (CR > 3.0) (Mosier & Mcquitty, 1940). Furthermore, item analysis extended beyond statistical metrics to incorporate theoretical and practical considerations (Hu, Leeuwen & Li, 2019). Acceptable Cronbach’s alpha values were ≥0.7 (Streiner, Norman & Health, 2015). Test-retest reliability was evaluated using the intraclass correlation coefficient (ICC), based on data from 50 AC patients who completed the questionnaire battery again two weeks later.

Content validity was assessed using the content validity index (CVI), rated by a panel of experts—including three palliative care specialists, one linguist, and one psychologist—on a 4-point Likert scale from 1 (strongly disagree) to 4 (strongly agree). Item-CVI values exceeding 0.7 indicated satisfactory content validity (Grant & Davis, 1997).

Patients’ physical symptoms and emotional distress have been identified as major predictors of cancer patients’ needs (Wang et al., 2018; Wang et al., 2021). Additionally, emotional distress is associated with reduced preferences for PC due to misconceptions and delayed gratification (Gerhart et al., 2016). This explains the selection of the DT and ESAS for evaluating concurrent validity in the original PPCI. Similarly, concurrent validity of the C-PPCI was examined using Pearson’s correlation coefficients between the C-PPCI and both the DT and ESAS. Construct validity of the C-PPCI was evaluated through exploratory factor analysis (EFA). Prior to EFA, sampling adequacy was confirmed using the Kaiser-Meyer-Olkin (KMO) test (>0.6) and Bartlett’s test of sphericity (p < 0.001). Factors were extracted based on eigenvalues > 1.0 and factor loadings > 0.40 (Tabachnick & Fidell, 2007). Confirmatory factor analysis (CFA) was performed using Mplus version 7.0, with acceptable model fit criteria including χ²/df ≤ 3, root mean square error of approximation (RMSEA) ≤ 0.08, comparative fit index (CFI) > 0.90, standardized root mean square residual (SRMR) ≤ 0.08, and Tucker-Lewis index (TLI) > 0.90 (Hu & Bentler, 1999).

Demographic characteristics

Of the 537 recruited participants, 444 (82.6%) completed the questionnaire. Their demographic characteristics are summarized in Table 1. The majority of these AC patients were female (53.8%) and married (93.5%), with lung cancer (27.5%) and breast cancer (15.8%) being the most common diagnoses. The subsamples for EFA and CFA exhibited similar demographic profiles (Table 1). Although not precisely measured, the average completion time ranged from 10 to 20 min, suggesting the feasibility of implementing the C-PPCI among AC patients.

Instrument modification

Based on expert recommendations during the translation and cultural adaptation phase, several modifications were made to enhance the clarity of the C-PPCI items. The panel noted that the PPCI assesses not only emotional and cognitive reactions of advanced cancer (AC) patients upon first learning about palliative care (PC) but also their needs. Consequently, the instrument’s name was revised to “Perceptions of Palliative Care and Needs Instrument” in Chinese to better reflect its scope. For Item 18, experts found the term “strangers” ambiguous and revised it to “strangers associated with providing palliative care” for clarity. To broaden the applicability of the C-PPCI to other terminally ill patients beyond those with AC, Item 9 was modified from “Think my cancer is out of control” to “Think my disease is out of control”. Additionally, since 2016, the terms “hospice”, “palliative care”, and “end-of-life care” have been officially standardized as “Anning Liaohu” in China (Ning, 2018); thus, “palliative care” was translated accordingly in the C-PPCI.

Item analysis

Table 2 presents the item discrimination results, with critical values exceeding 3.0 for all items. As shown in Table 3, item-total correlations were statistically significant across all items. However, items 2, 4, 16, 17, 20, 21, and 29 exhibited weak correlations. These items were retained in the C-PPCI as their item-total correlations approximated 0.30. Item 18 was excluded due to an item-total correlation substantially below 0.30. Item 7 was retained, despite its lower correlation, to maintain the three items assessing positive emotional reactions in the original PPCI.

Content validity

As presented in Table 3, the scale-level content validity index (S-CVI) for the C-PPCI was 0.99. The item-level content validity index (I-CVI) was 1.0 for all items except items 18 and 23, which had an I-CVI of 0.80. All CVI values met or exceeded 0.8, indicating strong content relevance for the C-PPCI (Grant & Davis, 1997).

Construct validity

Consistent with the original PPCI, we employed the same analytical approach to assess the construct validity of each domain in the C-PPCI separately (Milne et al., 2013). The data were suitable for principal component analysis, as evidenced by Kaiser-Meyer-Olkin (KMO) values ranging from 0.660 to 0.860 across the four domains and a significant Bartlett’s test of sphericity (p < 0.001). Similar to the original PPCI, exploratory factor analysis (EFA) identified four domains—emotional reactions, cognitive reactions, palliative care needs, and perceptions of burden—explaining 64% to 75% of the total variance per domain, as shown in Table 4. EFA results revealed nine factors with eigenvalues greater than 1.0 and factor loadings exceeding 0.4. Notably, the palliative care needs domain diverged from the original PPCI based on statistical criteria.

Specifically, items 22, 23, and 32, related to physical management (e.g., physical care, showering, and pain management), formed a factor labeled “physical needs”, accounting for 18.21% of the variance in the PC needs domain. Items 24, 27, and 28, pertaining to emotional or spiritual support, were grouped as “emotional and spiritual needs”, explaining 20.33% of the variance. Items 29, 30, and 33, focused on discussions about the illness, were categorized as “communication needs”, contributing 17.07% to the variance. Lastly, items 26, 31, and 34, addressing preparations for the future or conversations about dying, were labeled “preparation needs for the future”, accounting for 16.15% of the variance. Beyond statistical criteria, logical judgment led to the removal of item 25, as it was unrelated to other items in the emotional and spiritual needs factor.

As presented in Table 5, confirmatory factor analysis (CFA) for each subscale confirmed acceptable fit indices: χ²/df ranged from 0.000 to 2.289, RMSEA from 0.000 to 0.076, CFI from 0.941 to 1.000, TLI from 0.922 to 1.000, and SRMR from 0.000 to 0.055. Improved fit indices supported the removal of item 25 (Table 5). Overall, CFA validated a nine-factor model with robust fit indices: χ²/df = 1.574, RMSEA = 0.051, CFI = 0.917, TLI = 0.904, and SRMR = 0.072.

Concurrent validity

As shown in Table 6, significant positive correlations were observed between the DT and emotional reactions (r = 0.104, p = 0.029), palliative care needs (r = 0.117, p = 0.014), and perceptions of burden (r = 0.267, p < 0.001). However, no significant correlation was found between the DT and cognitive reactions (r = 0.000, p = 0.996). Similarly, positive correlations were identified between the ESAS and emotional reactions (r = 0.107, p = 0.024), palliative care needs (r = 0.279, p < 0.001), and perceptions of burden (r = 0.407, p < 0.001). No significant association was detected between ESAS and cognitive reactions (r = 0.064, p = 0.307).

Internal consistency reliability, test-retest reliability

As presented in Table 7, the Cronbach’s alpha coefficient for the total C-PPCI was 0.852, with values ranging from 0.820 to 0.872 for the four subscales, indicating robust internal consistency reliability. To evaluate test-retest reliability, 50 patients with AC completed the C-PPCI again after two weeks. The intraclass correlation coefficient (ICC) was 0.855 for the total scale and ranged from 0.751 to 0.815 for the subscales, demonstrating strong stability over time.

Strength and limitations

A key strength of this study is that the C-PPCI was established as a valid, reliable, and multifaceted tool for assessing AC patients’ perceptions, needs, and self-perceived burdens related to PC, using a concise set of items. It is particularly well-suited for patients encountering PC for the first time. Additionally, the development of the C-PPCI addresses a critical gap by providing a robust measure tailored to the Chinese cultural context, while also facilitating cross-cultural comparisons.

The primary limitation is that participants were recruited from a single center in Henan without integrated PC services. Chosich reported that cancer patients exposed to PC or familiar with it tend to harbor fewer negative emotions toward it (Chosich et al., 2020). Future studies should validate the C-PPCI in patients who have received PC services, encompassing both cancer and other life-threatening chronic conditions. PC services remain incompletely integrated into China’s mainstream healthcare system (Willemsen et al., 2021), and while home-based PC services are currently limited yet in high demand (Liu et al., 2021), future research could explore the C-PPCI’s application in home-based settings to promptly identify and address barriers to PC utilization. Furthermore, the integration of artificial intelligence (AI) and machine learning (ML) presents a promising frontier for enhancing patient-centered palliative care. AI-driven tools could potentially analyze patient-reported data from instruments like the C-PPCI in real-time to predict unmet needs, personalize symptom management plans, and facilitate more timely and effective communication between patients and providers (Dixon et al., 2024). Addressing these human-centered challenges through technology could help scale the benefits of palliative care, a crucial consideration in a large and resource-diverse country like China. For example, ML models could identify patient clusters with specific perception and need profiles, allowing for targeted educational interventions (Thacharodi et al., 2024). The C-PPCI could serve as a valuable data source for developing and validating such intelligent healthcare systems in the future.