Abstract
Introduction
The Serious Illness Conversation Guide (SICG) is a structured communication tool that equips clinicians with suggested language to facilitate serious illness conversations (SICs) with their patients. With prompts and questions on illness understanding, informational preferences, priorities, and care preferences, among others, the Guide facilitates dialogue that has been shown to improve communication, enhance patients’ emotional well-being, and reduce burdensome end-of-life interventions. 1–5
Originally developed in the United States and used primarily in Western, Anglophone settings, 6 the SICG was previously adapted for English-speaking patients in Singapore. 7 Yet many older adults in Singapore speak Mandarin, the first language of a quarter of the resident population. 8,9 While there exist other Chinese-language versions of the SICG, such as those developed by Stanford Medicine and Vancouver Coastal Health, 10,11 their lexical choices and styles may reflect linguistic norms specific to North American diaspora communities, and may not align with local usage or communication styles.
As language and cultural barriers are a documented obstacle to initiating goals-of-care discussion, 12–14 adaptations are essential to render the Guide contextually appropriate for clinical implementation. To that end, this study translated and iteratively refined a Mandarin-language version of the SICG for use in Singapore.
Objectives
This study aims to:
Document and explain the revisions introduced to the Mandarin-language SICG; Test the acceptability of the finalized Mandarin-language SICG in the cultural context of Singapore.
Methods
Settings and participants
Participants were recruited from outpatient clinics at the National Heart Center Singapore (NHCS) and the National Cancer Center Singapore (NCCS). Eligible patients were ≥21 years of age or older, aware of their diagnosis, had advanced cancer (stage III-IV) or heart failure (stage C-D), and could converse in Mandarin. Patients who lacked decision-making capacity or could not provide informed consent were excluded.
Purposive sampling was employed to ensure variation in age, gender, educational background, and diagnosis (cancer or heart failure). Recruitment was conducted by A.N. between September 2024 and April 2025. Sampling continued until thematic saturation was reached in Phase 2 and until 50 patients were recruited in Phase 3, providing adequate power for Wilcoxon signed-rank tests, assuming a moderate effect size (P(X > Y) = 0.67), α = 0.05, and power = 0.80.
Study design and procedures
This three-phase qualitative study involved (1) translation of the SICG, (2) two rounds of interview-based testing and revision, and (3) a final survey-based acceptability assessment. We consulted the Standards for Reporting Qualitative Research guidelines when preparing this article (see Supplementary Appendix SA1 for the annotated checklist). 15
Phase 1: SICG translation
A Mandarin version of the locally adapted English SICG was developed through forward translation by A.N., S.H.S.N., and Wang Xirui, who are fluent in both English and Mandarin. An independent researcher (Ke Yu) reviewed the draft for clarity and fidelity, and back-translated it into English. Discrepancies were resolved by discussion (A.N., S.H.S.N., Wang Xirui).
Phase 2: Iterative testing and revision
Two rounds of interviews were conducted. In Round 1 (September–October 2024), we recruited 8 cancer patients for structured interviews to obtain preliminary directions for revisions. In Round 2 (November 2024—January 2025), 17 patients with heart failure or cancer were recruited to reach thematic saturation.
Feedback from Round 1 informed revisions tested in Round 2. Participants assessed each Guide item as if it were posed by a clinician, rating each item on a 3-point scale (1 = “Unacceptable”; 2 = “Acceptable with changes”; 3 = “Acceptable as is”) and elaborating on its clarity, tone, and perceived relevance. Where two versions of the Guide were provided (Round 2), they were presented in a fixed sequence, and participants were instructed to evaluate each independently to minimize contrast bias. Interviews were audio-recorded, transcribed verbatim, and analyzed thematically.
Revisions were proposed independently by A.N. and S.H.S.N. and finalized through consensus discussions. As clinicians experienced in serious illness communication, S.H.S.N. and G.L.G.P. contributed emic/insider insights into prevailing communication norms. A.N., a nonclinician researcher, provided an etic/outsider perspective on clinical discourse. The combination of standpoints supported reflexive consideration of assumptions during Guide revision.
Phase 3: Final acceptability assessment
In Phase 3, we assessed the acceptability of the finalized Guide and tested whether revisions improved cultural fit through a survey instrument that compared earlier and revised versions for each item. Where Version 1 phrasing was reinstated, Version 3 was compared with its Version 2 counterpart. Participants were alternately assigned upon recruitment to complete one of two versions of the questionnaire (A or B), which differed in the order of Version presentation to mitigate order bias (Supplementary Appendix SA2).
Surveys were self-administered or conducted verbally by the recruiter for participants with literacy or vision difficulties. Each item was rated using a 4-point Likert scale (1 = “Very Unacceptable” to 4 = “Very Acceptable”). Data were collected over five weeks (March–April 2025) from 51 participants (26 with cancer; 25 with heart failure). Summary statistics were compiled in Excel, and Wilcoxon signed-rank tests were performed in R (v4.2.0).
Results
A total of 76 participants were recruited for this study (Table 1). They varied in age (mean 64 years, SD 11.03), gender (46.1% female), religion (Buddhism, Christianity, free-thinker), and education (nearly half with secondary or higher). Clinical diversity was also evident: 41 had cancer (26 Stage IV) and 35 had heart failure (33 Stage C). Most cancer patients were functionally well (Eastern Cooperative Oncology Group [ECOG] 0–1), and most heart failure patients had milder symptoms (New York Heart Association/NYHA I-II), though their responses revealed differing levels of acceptance, readiness, and comfort with SICs.
Participant Characteristics (n = 76)
Eastern Cooperative Oncology Group (ECOG) Performance Status Scale is widely adopted in cancer research to assess a patient’s ability to perform daily activities. As none of the participants fall under Grade 4 (“completely disabled”) or 5 (“dead”), these grades have been omitted from the table.
Fully active, able to carry out all predisease performance without restriction.
Restricted in physically strenuous activity but ambulatory and able to carry out work of a light or sedentary nature, e.g., light housework, office work.
Ambulatory and capable of all self-care but unable to carry out any work activities. Up and about more than 50% of waking hours.
Capable of only limited self-care, confined to bed or chair more than 50% of waking hours.
The New York Heart Association (NYHA) classification system is used to assess the severity of heart failure with reference to the functional limitations and symptoms experienced by patients. As none of the participants fall under Class IV (“Symptoms of heart failure at rest”), this was omitted from the table above.
No limitation of physical activity. Ordinary physical activity does not cause undue fatigue, palpitation, or shortness of breath.
Slight limitation of physical activity. Comfortable at rest. Ordinary physical activity results in fatigue, palpitation, shortness of breath, or chest pain.
Marked limitation of physical activity. Comfortable at rest. Less than ordinary activity causes fatigue, palpitation, shortness of breath, or chest pain.
Rating tabulation for Round 1 (phase 2)
Each participant evaluated 23 items from Version 1 of the Mandarin SICG. Table 2 highlights the 13 items that were rated “Acceptable as is” (rating = 3) by 6 or fewer participants in Round 1 (n = 8). Four (items 1–3, 1–16, 1–8, 1–21) received “Unacceptable” (rating = 1) from one participant; item 1–13 was rated “Unacceptable” by two participants. These ratings directed our attention to items in need of rephrasing or reframing. (Detailed results in Supplementary Table SA3.1 in Supplementary Appendix SA3, and Mandarin-language Guide items in Supplementary Appendix SA4).
Results from Round 1 of Phase 2 (n = 8)—Items Requiring Revision
Qualitative findings from Round 1 (phase 2)
Participants felt that some items did not reflect how clinicians typically communicated, and several interpreted them as signaling terminal decline. Despite being told that SICs were not limited to end-of-life care, five participants viewed questions about favorite activities (items 1–14) and current priorities (items 1–10) as implying that they had a short prognosis. These reactions may have been influenced by a limited understanding of their condition, as seen in responses to items 1–4 (“What do you understand about your condition?”).
Participants also reported discomfort with references to prognosis (items 1–7, 1–8) and potential decline (items 1–9, 1–11, 1–16). A common preference emerged for a gentler, more supportive tone and reassurance from clinicians where appropriate.
SICG revisions introduced in Round 1 (phase 2)
In response to these findings, Version 1 of the SICG underwent substantial revision. These revisions were guided by three aims:
To reassure patients that these questions do not imply imminent decline; To align the Guide’s language with phrases familiar to patients; To clarify clinicians’ intentions in raising potentially sensitive topics.
Table 3 provides a side-by-side comparison of Versions 1 and 2 of the Guide, with Guide items back-translated into English. Underlined text marks revisions discussed below.
Side-by-Side View of Versions 1 and 2 of Serious Illness Conversation Guide (English Back-Translated)
Reassurance and softening of prognostic language
Items 1–9 (functional decline) and 1–7 (prognostic uncertainty) were merged and reframed to emphasize clinician support:
“我会尽力帮您保持良好状态…我们在应付病情的当儿…”
[“I will do my best to help you stay well… as we manage your condition…”]
Item 1–8 (remaining time) was softened with a question that elicits the patient’s informational preferences, offering them the opportunity to decline receiving prognostic information:
“能知道自己有多少时间, 对您来说重要吗？”
[“Is it important to you to know how much time you have?”]
Cultural and linguistic adaptation
Terms were adapted for clarity and cultural fit. For example, “sources of strength” (item 1–13) was revised to sources of “moral support [精神上的支持]” to better reflect social-relational understandings of strength and resilience. The original term “strength [力量]” had in some instances, led to confusion or silence, suggesting poor cultural fit.
Clarification of intent
Item 2–16 was prefaced with a rationale to underscore the clinician’s concern for the patient’s well-being, helping build trust and rapport:
“能延长生命的医疗措施也可能会影响到病人与家属的生活素质, 也可能带来疼痛不便、心理上的压力、等等。”
[“Life-prolonging treatments may affect the quality of life of the patient and family members, and may also cause pain, inconvenience, psychological stress, etc.”]
An explanatory preamble was also added to the Guide as item 2–9:
“我接下会问您一些问题, 目的是了解您的价值观/您重视的是什么, 以确保您的治疗能更好地符合您的需求。”
[“I will be asking you some questions to understand your values/what is important to you, so as to ensure that your treatment can better meet your needs.”]
Rating tabulation for Round 2 (phase 2)
Round 2 asked participants to review and comment on both the original (Version 1) and revised (Version 2) Mandarin-language SICG items. Of the 21 items assessed, 15 were revised and directly compared. The remaining six were unchanged and received unanimous ratings of 3 (“Acceptable as is”).
Acceptability, defined by the number of participants giving a rating of 3 (“Acceptable as is”), improved for 12 of the 15 revised items. The greatest improvements were seen in items 2–8, 2–7, and 2–20, which saw an additional 9 participants giving a rating of 3. The smallest improvement was in item 2–13, for which an additional two participants gave a rating of 3, likely because the original wording (item 1–13) had already been deemed acceptable by 10 participants in Round 1. Items 2–10 and 2–17 showed no change, while item 2–6 declined in acceptability: only 6 participants rated the revised version “Acceptable as is”, compared with 14 for the original (item 1–6). (See Supplementary Table SA3.2).
Qualitative findings from Round 2
Revisions to prognostic items (items 2–8 and 2–7) were generally well-received and showed the greatest improvement in acceptability. Nonetheless, some participants still found the statements disheartening and preferred an emotionally sensitive tone that was less suggestive of the need to engage in end-of-life discussions. This sensitivity extended to items 2–10 (patient’s priorities). While some welcomed the use of examples (e.g., “family, continuing work, enjoying life”), others felt these implied that their condition was terminal.
Surprisingly, items 2–6 suffered a decline in acceptability following a revision based on feedback from participants about the involvement of family members. Version 1 asked if the clinician could share their understanding of the patient’s condition, while Version 2 included a follow-up question about relaying the information to the patient’s family members. Three participants rated it “Unacceptable” (rating = 1) due to concerns about worrying their loved ones, preferring instead to shield them from knowledge of their illness.
SICG revisions introduced in Round 2
Further revisions to Version 2 of the Guide aimed to:
Reassure patients that their condition is currently stable/under control; Emphasize the clinician’s concern for the patient’s quality of life; Simplify language to reduce cognitive burden and attentional fatigue.
Table 4 shows an item-level comparison between Versions 2 and 3. Underlining marks illustrative revisions described below.
Side-by-Side View of Versions 2 and 3 of Serious Illness Conversation Guide (English Back-Translated)
Reinforcing reassurance
In items 3–8 (remaining time), the question about informational preferences was reframed as a hypothetical:
“如果有一天我发现您的情况恶化, 您会希望我和您分享您有多少时间吗？”
[If I notice that your condition is getting worse one day, would you like me to share how much time you have?”]
This phrasing implicitly affirms the patient’s current stability and is intended to provide a gentler entry point for clinicians to explore their patients’ informational preferences regarding prognostic disclosures.
Emphasizing concern for quality of life
The rationale for discussing trade-offs was revised in items 3–16 to reinforce the clinician’s concern for the patient’s well-being. Couching the trade-offs as hypotheticals was expected to help the conversation feel less alarming for patients as well.
“因为我关注的是您的生活素质, 我想与您一起探讨一个假设性的问题…”
[“As I prioritise your quality of life, I wish to explore a hypothetical question with you…”]
Simplifying expressions
Item 3–20 was shortened to simply state the clinician’s recommendation, “I recommend that we ___. Is this ok?” The explanation in items 2–20 was left out:
“这样我们的治疗方案能让您继续做您最重视的事。”
[“In this way, our treatment plan will allow you to continue doing the things you value most.”]
This simplification reduces the cognitive load on patients and avoids inadvertently reaffirming unrealistic goals that some patients may have, such as remission or returning to a premorbid functional state.
Acceptability assessment in phase 3
Having finalized Version 3 of the Guide by the end of Phase 2, we conducted a final acceptability assessment. Participants rated two versions of each item using a 4-point Likert scale (1= “Very unacceptable” to 4= “Very acceptable”) (Supplementary Appendix SA2).
With no established benchmarks for acceptability, we defined an item as acceptable if (i) ≥75% of participants rated it 4 (“Very acceptable”), and (ii) ≥95% of participants rated it 3 (“Acceptable”) or 4. All Version 3 items met these thresholds except 3–10 (priorities) and 3–17 (family/social support), which satisfied criterion 2 but not 1, likely reflecting minimal differences from their Version 2 counterparts. Overall, findings indicate that Version 3 was well-received and suitable for clinical implementation (detailed results in Supplementary Table SA3.3).
By contrast, only three items from earlier versions of the SICG satisfied criterion 2 (items 1–15, 2–17, 1–23); none met criterion 1. Wilcoxon signed-rank tests were performed for each pairwise comparison. Of 15 paired items, 13 showed statistically significant improvements favoring Version 3 (p < 0.05, most at p < 0.001, see Table 5). Item 3–11 (future priorities) achieved a Wilcoxon V of 0, indicating unanimous preference for the revised version. Full p values are reported in Supplementary Table SA3.4.
Changes in Acceptability in Phase 3 (n = 51)
Discussion
Through iterative patient testing, we developed a culturally appropriate Mandarin-language adaptation of the SICG for use in Singapore (Table 4; Supplementary Appendix SA5). In Phase 3, every item received a median rating of “Very acceptable (rating = 4), with 19 of 21 items meeting study-defined acceptability criteria. Wilcoxon signed-rank tests showed statistically significant preference for 13 of the 15 revised items, supporting the Guide’s clinical and training applicability.
Across both rounds in Phase 2, participants expressed discomfort with items assessing their understanding of their condition (items 1–4, 2–4). These questions often provoked anxieties about undisclosed bad news or highlighted informational gaps, consistent with prior studies that show limited illness understanding among patients with cancer and heart failure. 16–19 Clinicians may therefore need to initiate SICs early and foster a climate of transparency in patient interactions.
Despite being translated from a locally adapted English-language SICG, items on prognosis and future care elicited considerable anxiety, underscoring how culturally sensitive topics may be interpreted differently across culturally diverse communities. While participants articulated diverse illness experiences and coping styles, most expressed discomfort or hesitancy about reflecting on future care needs and goals (items 1–11, 1–16, 2–11, 2–16, 3–11, 3–16). These reservations appeared unrelated to biomedical staging (e.g., disease stage, NYHA, ECOG), reinforcing the importance that clinicians approach goals-of-care discussions with tact and emotional sensitivity.
Barriers identified corroborated findings in the existing literature: the emotional burden of discussing serious illness, a desire to maintain optimism, and fears that engaging in goals-of-care conversations may hasten decline. 20–22 Even so, most of the participants claimed to be open to SICs, trusting their clinicians to have good reasons for initiating such conversations. Interestingly, many participants preferred to participate in these discussions themselves to avoid burdening their family members, diverging from the literature, which describes strong family-centered decision-making styles in Asian contexts. 23–25
This study has several limitations. First, participants in Phase 2 Round 2 and Phase 3 compared two versions of the same item, introducing potential contrast bias despite instructions to evaluate each item independently. Second, the Phase 3 sample size (n = 51) was not powered to identify small effects or allow detailed subgroup analyses. Finally, the scope was limited to Mandarin-speaking patients with cancer or heart failure. Although many participants had comorbidities and responded similarly to those without, findings should still be interpreted as indicative rather than definitive.
While Singapore’s multilingual environment may constrain generalizability, the findings remain relevant to other contexts where Mandarin is commonly spoken alongside other languages. Importantly, this study was designed to assess cultural and linguistic acceptability, and not to validate a new psychometric instrument. Future work may involve extending adaptation to Malay and Tamil speakers and evaluating the Mandarin-language SICG in clinical practice.
Conclusion
This study demonstrates the acceptability and cultural appropriateness of a Mandarin-language adaptation of the SICG for use with seriously ill patients in Singapore and potentially in the region. Iterative testing revealed robust support and a clear preference for the final version of the Guide (Supplementary Appendix SA5). Further research is warranted to explore the Guide’s regional applicability with other Mandarin-speaking communities.
Footnotes
Acknowledgments
The authors extend their gratitude to LCPC and the Lien Foundation for their generous support for this project, Dr Ke Yu for her assistance with reviewing the translation of the SICG, Miss Wang Xirui for her support in translation and patient recruitment, and its clinician partners (Professor Ravindran Kanesvaran, Dr Nagavalli, Dr Bernard Chua, Dr Chiang Jianbang, Associate Professor David Sim, Dr Louis Teo, Dr Shirlynn Ho, Dr Jason Chan, Dr Ryan Tan, Dr Stephanie Saw, Dr Teh Yi-Lin, Dr Koo Si Lin) from NCCS and NHCS for their support.
Authors’ Contributions
A.N.: Investigation, formal analysis, data curation, writing—original draft, writing—review and editing, and project administration. G.L.G.P.: Writing—validation, review, and editing. S.H.S.N.: Conceptualization, methodology, supervision, writing—review and editing, and funding acquisition.
Data Availability Statement
The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.
Ethics Approval
This study was approved by the institutional review board at the Singapore Centralised Institutional Review Board (CIRB) (approval no.: 2022/2736) on Feb 14, 2023. All participants gave written consent prior to study participation (interviews or survey completion).
Author Disclosure Statement
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding Information
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Lien Center for Palliative Care Research Award [Grant number: LCPC-EX22-0001].
Appendix
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
