Development of a Nursing Quality Evaluation Indicator System for Lung Cancer in Traditional Chinese Medicine Based on the Donabedian Model: A Delphi Study Authors: Ting Duan, Yuxin Liu, Huihua Wang Published: 2026 ABSTRACT AbstractBackgroundCurrently, no unified quality evaluation standard exists specifically for Traditional Chinese Medicine (TCM) nursing in lung cancer in China, and existing indicators do not capture TCM’s unique features. To address this gap, we aimed to construct a preliminary scientific and disease specific quality evaluation framework for Traditional Chinese Medicine (TCM) nursing in lung cancer patients based on expert consensus, which can serve as a theoretical foundation for standardized management and future empirical validation.MethodsGuided by Donabedian’s structure process outcome model, we developed an initial indicator pool through a systematic literature review and semi structured interviews. A three round Delphi consultation was conducted with 17 national experts. Indicators were screened using importance/feasibility scores, coefficients of variation, and full-score rates. An analytic hierarchy process (AHP) was used to assign weights. No clinical validation has been performed at this stage.ResultsAll three rounds achieved a response rate of 100%. The expert authority coefficients increased from 0.853 to 0.894, and Kendall’s W reached 0.304 (all P < 0.01). The final framework comprised three primary, 17 secondary, and 45 tertiary indicators, with coefficients of variation <0.25 and consistency ratios <0.10. The three primary domains were structure (focusing on staffing, training, and systems), process (emphasizing symptom management, TCM techniques, and safety), and outcome (covering symptom control and TCM syndrome improvement).ConclusionWe developed a consensus‑based quality evaluation framework for TCM nursing in lung cancer that demonstrated good content validity and expert agreement. However, these findings are based solely on expert opinion; their predictive validity, clinical feasibility, and impact on patient outcomes require prospective multicenter validation before they can be recommended for routine practice. This framework provides a theoretical reference for future studies.Keywords: lung cancer, traditional Chinese medicine nursing, quality evaluation, Delphi method, analytic hierarchy process, Donabedian model Introduction Lung cancer ranks first in terms of incidence and mortality among malignant tumors worldwide.1–3Patients often experience cough, hemoptysis, dyspnea, and cancer related fatigue, with the prognosis heavily dependent on nursing quality. Traditional Chinese Medicine (TCM) nursing has become an essential part of comprehensive lung cancer management, effectively reducing the side-effects of chemoradiotherapy, relieving symptoms, and regulating immunity.4–6However, China currently lacks a unified quality evaluation standard, specifically for TCM nursing in lung cancer. Existing assessments predominantly focus on process indicators while neglecting outcomes, leading to wide variations in practice and hindering systematic development.7–10 Current general TCM nursing quality indicators emphasize basic TCM technique execution but do not capture key TCM-specific outcomes such as syndrome evolution or the integration of syndrome differentiation into symptom management. These are critical for lung cancer patients, whose symptom burden fluctuates with chemotherapy cycles and requires dynamic syndrome-based adjustments. This framework adds value by providing a standardized, measurable structure that enables future empirical validation and multi-center comparison, which are currently lacking. Donabedian’s structure process outcome (SPO) model is a classic framework widely used in nursing quality research.11–13While general TCM nursing quality indicators exist, none is specifically tailored to the complex symptom profile and prolonged care needs of patients with lung cancer. The Delphi technique has been increasingly applied in nursing research to develop consensus-based indicators.14,15 Therefore, this study aimed to construct a specialized evaluation system using the SPO model, integrating literature review, semi structured interviews, the Delphi method, and an analytic hierarchy process (AHP) to provide a theoretical reference for objective quality assessment and continuous improvement. We acknowledge that this is a consensus based developmental study and subsequent clinical validation is required. Methods Study Design A Delphi study was conducted and reported according to the Conducting and Reporting DElphi Studies (CREDES) guidelines,16 and the methodological approach was further explained by established Delphi reporting standards.17 Construction of the Preliminary Indicator System The research team consisted of seven members: one deputy nursing director, two head nurses, one clinical nurse, one physician, and two graduate students who were responsible for literature retrieval, indicator drafting, expert selection, data analysis, and weight calculation.18 Systematic Literature Review A systematic search was performed from inception to November 30, 2025, in CNKI, Wanfang, VIP, PubMed, and Embase using the search terms (“lung cancer” OR “pulmonary neoplasm”) AND (“TCM nursing” OR “traditional Chinese medicine nursing” OR “syndrome differentiation nursing”) AND (“quality evaluation” OR “quality indicator” OR “structure‑process‑outcome”). Studies were included if they addressed nursing quality evaluation in cancer or chronic diseases; involved TCM interventions or nursing management; were guidelines, expert consensus, systematic reviews, or original research; and were published in Chinese or English. Conference abstracts, editorials, case reports, and studies without clear definitions of indicators were excluded. Two researchers independently screened the titles and abstracts, and any disagreements were resolved by a third reviewer. The same two researchers independently extracted quality indicators from the eligible full texts using content analysis, categorizing them into structure, process, and outcome domains; indicators mentioned in at least two sources were retained for the preliminary pool. Semi Structured Interviews From December 2025 to March 2026, separate from the Delphi expert panel, we conducted face to face semi structured interviews with a purposive sample of TCM nursing experts (n = 5), clinical nurses (n = 7), and lung cancer patients (n = 8) from a tertiary hospital in Jiangxi Province. The interview guide included core questions: “What are the most critical quality links in TCM nursing for lung cancer patients?”, “What are the current deficiencies in quality assessment?”, “What specific indicators do you think should be included?”, and “What outcomes matter most to patients?” The interviews were conducted in private rooms, lasting 30–45 minutes, audio recorded, and transcribed verbatim within 24 hours. Data saturation was reached after the 15th interview (no new themes emerged in the three consecutive interviews). Transcripts were independently coded by two researchers using Colaizzi’s seven step method, with intercoder agreement (Kappa = 0.86) achieved through consensus meetings. The research team then held a consensus meeting to merge the literature‑derived indicators with interview‑derived themes, resolving overlaps and phrasing discrepancies to form the initial Delphi questionnaire comprising 72 preliminary indicators. Delphi Expert Consultation Expert Panel Selection Experts were recruited from tertiary hospitals and medical universities across China. Inclusion criteria were as follows: at least 10 years of relevant work experience, a bachelor’s degree or higher, an intermediate or higher professional title, expertise in clinical nursing, nursing management, clinical medicine, or nursing education, and voluntary participation. Seventeen experts were enrolled.19 Delphi Process Three rounds of consultation were conducted via Email between December 2025 and May 2026. The Round 1 questionnaire contained three sections: study background, objectives, and instructions; a 5‑point Likert scale (1 = not important/feasible, 5 = very important/feasible) for each of the 72 indicators, with open ended comment boxes for each domain and general suggestions; and expert demographics, self rated familiarity (5‑point scale), and judgment basis (theoretical analysis, practical experience, literature, subjective judgment) with assigned scores. After each round, we provided a summary of the group statistics (mean, standard deviation, CV, full‑score rate, and percentage of experts recommending deletion) along with all anonymized qualitative comments, and experts were asked to reevaluate their scores based on this feedback. For subsequent rounds, indicators were deleted if they met any of the following criteria: ≥30% of experts recommended deletion, mean importance ≤3.5 combined with CV ≥0.25, or full‑score rate <20%.20 The consultation process ceased when the comment rate fell below 15% and Kendall’s W exceeded 0.3.21,22 Statistical Analysis and Weight Calculation Data were entered into Excel and analyzed using SPSS 26.0. Expert positivity was measured using response rate. The expert authority coefficient (Cr) was calculated as Cr = (Ca + Cs)/2, where Ca (judgment coefficient) was based on theoretical analysis (0.3), practical experience (0.5), literature (0.1), and subjective judgment (0.1), and Cs (familiarity) was self rated on a 5‑point scale (0.2–1.0); Cr >0.80 indicated high authority. Coordination was assessed using Kendall’s W and chisquare tests (P < 0.05, considered significant). We used the coefficient of variation (CV) as a supplementary indicator of expert consensus. While Likert-scale data lack a true zero, the CV provides a standardized measure of relative dispersion that facilitates comparison across indicators with different mean values and is a commonly reported metric in Delphi-based indicator studies. In this study, CV < 0.25 was used as one of the retention criteria, alongside mean importance scores and full-score rates, to ensure that only indicators with both high importance and acceptable consistency were retained. For the analytic hierarchy process (AHP), we used Yaahp 12.6 to calculate weights. We constructed pairwise comparison matrices based on the mean importance scores from the final Delphi round, converting score differences into Saaty’s 1–9 scale (eg, a difference of 1 point = ratio 3, difference of 2 points = ratio 5). All consistency ratios (CR) were required to be <0.10.23,24 When initial CR exceeded 0.10, we identified the largest residual in the pairwise comparison matrix and made the smallest possible adjustments consistent with the original expert scores, then rechecked CR. The final CR values reported are those after this iterative refinement. We acknowledge that this adjustment process may introduce some mathematical bias, and therefore the resulting weights should be interpreted as exploratory rather than definitive. Results General Information of Experts Seventeen experts participated in the study (Table 1). The median age was 46.5 years (range = 35–55 years), and median work experience was 20 years (range = 10–35 years). Most (94.12%) held senior titles and 70.59% had a master’s degree or above. Table 1General Information of Consulting Experts (N = 17)VariablesClassificationNumber of PeoplePercentage (%)Age35–45 years847.06>45–55 years952.94Education levelBachelor’s degree529.41Master’s degree635.29Doctorate degree or above635.29Working years10–20 years952.94>20–30 years529.41>30 years317.65Professional titleSupervisor nurse15.88Associate senior1058.82Senior635.29PositionClinical nurse529.41Head nurse529.41Director635.29Vice president15.88Work field*Clinical nursing1058.82Nursing management635.29Clinical medicine529.41Teaching or research847.06Note: * Work field is multiple choice. General Information of Consulting Experts (N = 17) Note: * Work field is multiple choice. Expert Positivity and Authority Coefficient All 17 questionnaires were returned during each round (100% response rate). The proportion of experts providing suggestions decreased from 64.71% (round 1) to 41.18% (round 2) and 11.76% (round 3), respectively. The expert authority coefficients (Cr) were 0.853, 0.874, and 0.894 across all rounds (all >0.80), indicating high credibility (Figure 1A). Changes of indicators in three‑round Delphi survey. (A) Expert suggestion rate (%) and Authority coefficient (Cr). (B) Kendall‘s W of expert consensus. Degree of Opinion Coordination Kendall’s W increased progressively from 0.180 to 0.304 in round 3 (all P < 0.01), indicating the convergence of expert opinions (Figure 1B and Table 2). Table 2Kendall’s Harmony Coefficient Values and Hypothesis Testing for Three Rounds of ConsultationsIndicator LevelRound 1Round 2Round 3Kendall’s Wχ2P valueKendall’s Wχ2P valueKendall’s Wχ2P valuePrimary indicators0.30310.316<0.010.37112.600<0.010.35312.000<0.01Secondary indicators0.13242.520<0.010.12341.974<0.010.30282.150<0.01Tertiary indicators0.158196.344<0.010.185349.047<0.010.316236.704<0.01Overall0.180294.562<0.010.189434.452<0.010.304331.092<0.01 Kendall’s Harmony Coefficient Values and Hypothesis Testing for Three Rounds of Consultations Final Indicator System and Weights After three rounds, 27 indicators were deleted or merged, leaving three primary, 17 secondary, and 45 tertiary indicators. All CVs were <0.25. The AHP derived weights for all indicators are presented in Table 3. All CRs were <0.10, indicating consistent pairwise judgment. Table 3Indicator and Weight of the Nursing Quality Evaluation System for TCM Nursing of Lung Cancer at All LevelsIndicatorImportance Assignment ()Coefficient of VariationRate of Maximum Scores (%)Weighted ValueI Structural quality4.94 ± 0.240.0594.12I-1 Human resource allocation4.94 ± 0.240.0594.120.110I-1-1 Proportion of nurses with ≥3 years of TCM nursing experience in oncology department4.88 ± 0.330.0788.240.037I-1-2 Certification rate of provincial-level or above TCM specialist nurses4.47 ± 0.510.1247.060.016I-1-3 Ratio of TCM nursing staff to actual open beds in oncology department4.94 ± 0.240.0594.120.057I-2 Training and education4.94 ± 0.240.0594.120.110I-2-1 Accuracy rate of TCM syndrome differentiation nursing assessment4.94 ± 0.240.0594.120.057I-2-2 Participation rate in TCM nursing research4.53 ± 0.510.1152.940.016I-2-3 Pass rate of TCM nursing technique operation examination4.88 ± 0.330.0788.240.037I-3 Technical support4.76 ± 0.440.0976.470.065I-3-1 Availability rate of lung cancer-specific TCM assessment scales4.41 ± 0.510.1141.180.016I-3-2 Information coverage rate of TCM nursing assessment4.82 ± 0.390.0882.350.049I-4 Equipment and supplies4.47 ± 0.510.1247.060.033I-4-1 Compliance rate of TCM characteristic nursing equipment configuration4.94 ± 0.240.0594.120.023I-4-2 Compliance rate of TCM nursing equipment maintenance4.41 ± 0.510.1141.180.005I-4-3 Pass rate of TCM nursing consumables quality4.41 ± 0.510.1141.180.005I-5 System and standards4.94 ± 0.240.0594.120.110I-5-1 Perfection rate of TCM nursing quality management system4.94 ± 0.240.0594.120.061I-5-2 Formulation rate of SOP for lung cancer TCM nursing techniques4.88 ± 0.330.0788.240.035I-5-3 Awareness rate of TCM nursing system4.41 ± 0.510.1141.180.014II Process quality4.94 ± 0.240.0594.12II-1 Symptom management4.94 ± 0.240.0594.120.084II-1-1 Accuracy rate of syndrome-based intervention for cancer pain4.88 ± 0.330.0788.240.010II-1-2 Accuracy rate of syndrome-based intervention for nausea and vomiting4.88 ± 0.330.0788.240.010II-1-3 Completion rate of TCM-specific assessment for cough, wheezing, and sputum4.94 ± 0.240.0594.120.021II-1-4 Compliance rate of standardized sputum evacuation nursing4.94 ± 0.240.0594.120.021II-1-5 Implementation rate of TCM exercise intervention for cancer-related fatigue4.94 ± 0.240.0594.120.022II-2 Emotional nursing4.88 ± 0.330.0788.240.055II-2-1 Implementation rate of syndrome-based music therapy4.82 ± 0.390.0882.350.018II-2-2 Guidance rate of mind-calming health exercises (Baduanjin, Tai Chi, etc.)4.94 ± 0.240.0594.120.037II-3 Dietary nursing4.47 ± 0.510.1247.060.025II-3-1 Accuracy rate of TCM constitution and syndrome differentiation before medicated diet intervention4.41 ± 0.510.1141.180.008II-3-2 Application rate of TCM strategies for treatment-related dietary problems4.47 ± 0.510.1247.060.017II-4 Application of TCM nursing techniques4.94 ± 0.240.0594.120.084II-4-1 Compliance rate of TCM nursing techniques (acupoint application, auricular seed pressing, moxibustion, cupping, etc.)4.94 ± 0.240.0594.120.024II-4-2 Application rate of TCM nursing techniques in perioperative period4.94 ± 0.240.0594.120.024II-4-3 Implementation rate of preventive TCM intervention before chemotherapy4.94 ± 0.240.0594.120.024II-4-4 Dynamic screening rate of contraindications for TCM nursing techniques4.88 ± 0.330.0788.240.012II-5 Syndrome differentiation assessment4.88 ± 0.330.0788.240.055II-5-1 Completeness rate of TCM syndrome differentiation assessment information4.94 ± 0.240.0594.120.016II-5-2 Implementation rate of psychological scale assessment (PHQ-9, GAD-7)4.94 ± 0.240.0594.120.016II-5-3 Accuracy rate of TCM syndrome differentiation for emotional state4.88 ± 0.330.0788.240.008II-5-4 Accuracy rate of TCM syndrome diagnosis for lung cancer4.94 ± 0.240.0594.120.015II-6 Rehabilitation nursing4.35 ± 0.490.1135.290.016II-6-1 Implementation rate of individualized TCM respiratory rehabilitation training for lung cancer4.35 ± 0.490.1135.290.016II-7 Health education4.47 ± 0.510.1247.060.025II-7-1 Implementation rate of individualized TCM characteristic health education plan for lung cancer4.47 ± 0.510.1247.060.025II-8 Safety management4.94 ± 0.240.0594.120.084II-8-1 Rate of TCM nursing risk assessment4.94 ± 0.240.0594.120.042II-8-2 Reporting rate of TCM nursing adverse events4.94 ± 0.240.0594.120.042III Outcome quality4.58 ± 0.510.1158.82III-1 Symptom control effect4.88 ± 0.330.0788.240.055III-1-1 Comprehensive improvement rate of lung cancer-related symptoms (cough, wheezing, sputum, etc.)4.94 ± 0.240.0594.120.022III-1-2 Mean reduction rate of Numerical Rating Scale (NRS) for cancer pain4.88 ± 0.330.0788.240.011III-1-3 Control rate of chemotherapy-related toxicities by TCM intervention4.94 ± 0.240.0594.120.022III-2 TCM syndrome improvement4.88 ± 0.330.0788.240.055III-2-1 Improvement rate of Qi deficiency syndrome4.88 ± 0.330.0788.240.010III-2-2 Improvement rate of Yin deficiency syndrome4.88 ± 0.330.0788.240.009III-2-3 Improvement rate of phlegm-dampness syndrome4.94 ± 0.240.0594.120.018III-2-4 Improvement rate of blood stasis syndrome4.94 ± 0.240.0594.120.018III-3 TCM nursing safety4.35 ± 0.490.1135.290.013III-3-1 Closed-loop rate of TCM nursing adverse events4.35 ± 0.490.1135.290.013III-4 Satisfaction4.41 ± 0.510.1141.180.020III-4-1 Patient satisfaction4.47 ± 0.510.1247.060.014III-4-2 Medical staff satisfaction4.41 ± 0.510.1141.180.006Note: The weighted value refers to the combined weight of the indicator in the overall system; the coefficients of variation for all indicators are <0.25, indicating good consistency.Abbreviations: SOP, standard operating procedures; NRS, Numerical Rating Scale; PHQ‑9, Patient Health Questionnaire‑9; GAD‑7, Generalized Anxiety Disorder‑7. Indicator and Weight of the Nursing Quality Evaluation System for TCM Nursing of Lung Cancer at All Levels Note: The weighted value refers to the combined weight of the indicator in the overall system; the coefficients of variation for all indicators are <0.25, indicating good consistency. Abbreviations: SOP, standard operating procedures; NRS, Numerical Rating Scale; PHQ‑9, Patient Health Questionnaire‑9; GAD‑7, Generalized Anxiety Disorder‑7. Discussion Scientific Rigor and Methodological Considerations We employed Donabedian’s SPO model as the theoretical framework, which has been widely validated in nursing quality research.25–27The combination of a systematic literature review, structured interviews, a three round Delphi, and AHP enhances the methodological rigor and content validity of the indicator system.28 Our expert panel was highly qualified (94.12% senior titles, 70.59% master’s degree or above). The 100% response rate, rising Cr values (>0.85), Kendall’s W reaching 0.304 (P < 0.01), and all CVs <0.25 collectively suggest strong expert consensus and acceptable content validity.29,30 However, it is critical to emphasize that these statistics reflect consensus reliability and content validity, not empirical criterion or clinical predictive validity. The three domains of the framework are not independent but form a logical chain: adequate structural resources (staffing, training, systems) enable high-quality process implementation (symptom management, TCM techniques, safety), which in turn drives favorable outcomes (symptom control, syndrome improvement, satisfaction). This interconnectedness ensures that the framework can be used not only for retrospective evaluation but also for prospective quality improvement—identifying structural gaps that may be contributing to suboptimal processes or outcomes. Interpretation of Weight Distribution Among the primary indicators, process (0.428) and structure (0.428) dominated, whereas outcome (0.144) carried less weight. This aligns with Donabedian logic but also reflects the experts’ belief that in the context of TCM nursing, rigorous process management is the most direct driver of quality, especially given the fluctuating clinical course of lung cancer.9,31 Collectively, these weighted indicators can serve as practical monitoring priorities for nursing managers, supporting targeted quality audit and continuous quality improvement for lung cancer-specific TCM nursing services. Among the process indicators, “symptom management”, “application of TCM techniques”, and “safety management” all had high weights (0.084), highlighting the prioritization of direct symptom relief (cough, fatigue, nausea) and risk prevention.32–35 Notably, guidance on active Daoyin exercises (Baduanjin) was weighted higher (0.037) than passive music therapy (0.018), suggesting a preference for active patient participation.36,37 In the structural domain, “human resources”, “training”, and “systems” each had a weight of 0.110, emphasizing that adequate staffing and competent syndrome differentiation are foundational.38 Interestingly, “certification rate of TCM specialist nurses” had a low weight (0.016), possibly because specialist nurse training in TCM is still evolving in China, and experts prioritize practical bedside competency over formal credentials. Outcome quality, though lowest in total weight, provides a dual‑axis evaluation combining Western symptom control (eg, NRS reduction) with TCM syndrome improvement (Qi, Yin, Phlegm‑Dampness, Blood Stasis). This dual approach is a key innovation, as it moves beyond superficial symptom counting to embrace TCM’s holistic “root and branch” philosophy.39 Clinical Relevance and Intended Application This system provides a theoretical reference for nurses to deliver standardized syndrome differentiation care and for managers to conduct quantitative monitoring. All indicators were derived from routine documentation or HIS systems, thus minimizing the extra workload. However, this system has not yet been tested in a live clinical environment. However, its clinical utility, sensitivity to change, and impact on patient outcomes remain unknown. Previous studies have emphasized that nurse led needs assessment can improve patient reported outcomes,40 but such benefits require prospective validation. We strongly caution against interpreting this system as a validated “clinical tool” at this stage; rather, it represents a well consolidated expert consensus framework. Limitations This study had several limitations. First, importance and feasibility were assessed using a single combined score in each round, which may have conflated two distinct constructs. Future studies should employ separate Likert scales for importance and feasibility to allow more refined indicator selection. Second, while the AHP matrix adjustment followed standard algorithms, this process may introduce mathematical bias, and the resulting weights should be considered exploratory rather than definitive. Third, all experts were recruited from tertiary teaching hospitals in mainland China, which may limit the generalizability of the indicator system to primary care settings and other healthcare systems. Fourth, the indicator system was purely consensus-based; we did not perform prospective clinical validation or correlation analysis with patient outcomes. Therefore, the terms “reliability” and “validity” used in this study refer strictly to content validity, not empirical performance. Fifth, although the indicators were designed based on routine documentation, the actual time burden for bedside nurses was not measured empirically. Future multicenter studies should prospectively apply this system to refine the weights and assess its clinical utility. Conclusion Using Donabedian’s SPO model, combined with a systematic literature review, semi structured interviews, a three round Delphi, and AHP, we successfully constructed a preliminary consensus‑based quality evaluation framework for TCM nursing for lung cancer patients. The final system comprises three primary, 17 secondary, and 45 tertiary indicators. Among the most important indicators are: in structure—human resource allocation, training and education, and systems and standards (each weight 0.110); in process—symptom management, application of TCM nursing techniques, and safety management (each weight 0.084); and in outcome—symptom control effect and TCM syndrome improvement (each weight 0.055). The framework can be used by nursing managers for baseline quality assessment, targeted improvement planning, and monitoring of changes over time. It demonstrates strong content validity and expert consensus, and reflects the unique characteristics of TCM nursing by integrating syndrome differentiation with symptom management. Nevertheless, this framework requires empirical clinical validation before it can be adopted for routine quality improvements. This study provides a theoretical foundation and clear direction for future prospective validation studies aimed at standardizing TCM nursing quality in China. Funding Statement This study was funded by the Science and Technology Program of the Jiangxi Administration of Traditional Chinese Medicine (grant number SZYYB20245320). Data Sharing Statement The datasets generated and/or analyzed during the current study are not publicly available because of the inclusion of sensitive interview transcripts and expert personal information, but are available from the corresponding author upon reasonable request. Ethics Approval Approved by the Ethics Committee of the Affiliated Hospital of Jinggangshan University (No. [2024]125). Informed Consent Prior to participation, written informed consent was obtained from all experts after a full disclosure of the study’s purpose, methods, potential risks, and data confidentiality measures. Participants were explicitly informed of their right to withdraw from the study at any time without penalty. Author Contributions All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Ting Duan contributed to the conception, study design, data acquisition, and drafting of the manuscript. Yuxin Liu contributed to the study supervision, final approval, and overall project administration. Huihua Wang contributed to the literature search, expert liaison, and manuscript revision. Disclosure The authors report no conflicts of interest in this work. References Associated Data Data Availability Statement The datasets generated and/or analyzed during the current study are not publicly available because of the inclusion of sensitive interview transcripts and expert personal information, but are available from the corresponding author upon reasonable request.