Experts Build Consensus Framework to Prevent Blood Clots in Lung Cancer Patients

A research team in Nanjing, China, has built a consensus-based service-delivery framework designed to translate widely endorsed clinical guidelines into concrete hospital workflows, staffing arrangements, and measurable quality indicators. The study, published in Frontiers in Cardiovascular Medicine, used a mixed-methods modified Delphi approach and produced a structure of 3 first-level domains, 11 second-level components, and 34 third-level indicators spanning structural capacity, care processes, and outcome monitoring.

The clinical problem the framework addresses is well established. Patients with lung cancer face an elevated risk of venous thromboembolism because of tumor-related hypercoagulability, anticancer treatment, hospitalization, reduced mobility, invasive procedures, and coexisting conditions. When clots do form, they can prolong hospital stays, increase healthcare costs, interrupt cancer treatment, and worsen prognosis. National and international guidelines already recommend routine risk assessment and appropriate prophylaxis for patients with cancer, but the study’s authors argue that guideline availability does not automatically produce consistent implementation at the bedside. Gaps can appear at many points: incomplete risk assessment, uncertainty about which professional is responsible for which step, insufficient patient education, limited medication safety monitoring, weak post-discharge continuity, and inadequate audit-and-feedback mechanisms.

To capture those organizational failures systematically, the team organized its work around Donabedian’s Structure–Process–Outcome model, a classic quality-improvement framework that links institutional capacity, the delivery of care, and measurable results. Candidate framework items were generated from two sources. The first was a structured literature review covering seven Chinese and English databases from inception to May 31, 2025. The search yielded 1,955 records; after duplicate removal, screening, and full-text assessment, 11 publications, including multiple international and Chinese guidelines on cancer-associated thrombosis, were included. The review was explicitly undertaken for item generation rather than effect estimation, and two researchers independently screened the literature, extracting information on risk assessment tools, prophylaxis processes, multidisciplinary roles, patient education, follow-up requirements, medication safety, and quality monitoring.

The second source was a series of one-to-one semi-structured interviews conducted between February and April 2025 at Nanjing Drum Tower Hospital. Using stratified purposive sampling, the researchers interviewed 16 healthcare professionals: 9 physicians, 3 nurses, 2 clinical pharmacists, and 2 hospital administrators. Interviews lasted roughly 30 to 45 minutes, were audio-recorded and transcribed verbatim, and were analyzed in NVivo 12.0 using the Consolidated Framework for Implementation Research as the initial coding scheme. Two researchers independently coded the transcripts, resolving disagreements through discussion and, where necessary, referral to a senior third researcher. Recruitment continued until successive interviews produced no substantially new categories.

The qualitative analysis identified facilitators in 7 CFIR constructs and barriers in 15. On the positive side, participants cited the low cost of prophylactic medications, supportive national guidelines and policies, leadership engagement, accessible training during departmental morning handovers, and clinicians’ genuine belief that prevention benefits patients. The barriers were more numerous and revealing: the complexity of managing medication adherence and monitoring drug interactions and adverse reactions, limited patient health literacy and awareness of thrombosis, gaps in clinicians’ own knowledge, unclear professional responsibilities, heavy workloads with no dedicated prevention staff, poor compatibility with existing workflows, ineffective communication feedback loops, punitive incentive structures, and the absence of standardized prevention workflows and integrated outcome data. Participants also noted that turnover among interns, postgraduate trainees, and rotating physicians undermined consistent implementation, and that lung-cancer-specific assessment tools were lacking.

These findings were consolidated into an initial Delphi questionnaire containing 3 first-level domains, 11 second-level components, and 38 candidate indicators. Fifteen experts, all from Nanjing Drum Tower Hospital, a tertiary hospital in Jiangsu Province, then rated every indicator on importance and feasibility using 5-point Likert scales across two rounds conducted between April and June 2025. The panel comprised 4 respiratory medicine specialists, 3 medical oncologists, 3 clinical pharmacists, 3 nursing administration experts, and 2 hospital administration specialists. Their mean age was 41.87 years, mean professional experience 16.67 years, twelve held master’s or doctoral degrees, and eleven held senior titles. All 15 returned valid questionnaires in both rounds, a 100 percent response rate, and the expert authority coefficient was high at 0.943 in round 1 and 0.953 in round 2.

Consensus statistics told a more nuanced story. Kendall’s coefficient of concordance ranged from 0.249 to 0.292 across rounds and rating dimensions, values the authors describe as indicating limited overall agreement despite statistically significant tests. In round 1, W was 0.292 for importance and 0.250 for feasibility; in round 2, 0.249 and 0.277 respectively. The researchers deliberately refrained from interpreting changes in W between rounds as improvement or deterioration, since the item set and wording differed. After round 1, one indicator was deleted, six were merged into three, and twelve were revised for clarity and reduced overlap. After round 2, all retained indicators were confirmed, yielding the final 34-item framework.

One of the study’s most practically useful findings is the systematic gap between what experts consider important and what they consider feasible. Mean importance and feasibility scores were generally high across the framework, but feasibility lagged for several implementation-dependent indicators. Three retained items fell below the predefined feasibility threshold of 4.00 in round 2: staff training and competency assessment (3.93), an electronic risk assessment and clinical decision-support system embedded in health records (3.93), and a structured post-discharge follow-up system for medication safety and self-monitoring (3.60). Multidisciplinary review and optimization of pharmacological prophylaxis also showed a notable gap. The research team retained these items as implementation priorities despite the lower feasibility scores, arguing that the pattern points to the true bottlenecks: organizational readiness, information system integration, staffing capacity, workflow redesign, and continuity-of-care infrastructure rather than any shortage of clinical conviction.

The final framework’s content maps directly onto hospital operations. Structural capacity indicators cover a standardized institutional prevention protocol across the care pathway, integration of prevention into hospital quality management, a multidisciplinary prevention team with clearly defined professional responsibilities, and resources including evidence-based pharmacological prophylaxis options, mechanical prophylaxis devices, and patient education materials in written and video formats. Care process indicators cover timely and repeated thrombosis and bleeding risk assessment using validated tools such as the Padua Prediction Score and the IMPROVE Bleeding Risk Score, standardized referral pathways for suspected deep vein thrombosis, escalation pathways for suspected pulmonary embolism, bedside alerts and daily handover verification for high-risk patients, daily lower-limb assessment, early mobilization support, screening for contraindications before mechanical prophylaxis, pre-prophylaxis medication history review, standardized nursing procedures for anticoagulant administration, admission-based patient education, discharge medication safety counseling, and structured post-discharge follow-up.

The outcome monitoring domain defines eight provisional measures with operational definitions, calculation formulas, and observation windows. These include VTE and bleeding risk assessment rates within 24 hours of admission, pharmacological, mechanical, and combined prophylaxis implementation rates among eligible patients, in-hospital incidence of objectively confirmed new deep vein thrombosis or pulmonary embolism, bleeding events among patients exposed to prophylactic anticoagulation using International Society on Thrombosis and Haemostasis definitions, and in-hospital VTE-related mortality. The authors are notably careful about interpretation: prophylaxis implementation rates are descriptive measures, not universal higher-is-better targets, and higher rates alone do not demonstrate more appropriate care, since both missed prophylaxis among eligible patients and potentially inappropriate prophylaxis must be considered. These additional specifications were developed after the Delphi consultation and require prospective testing and external measurement validation.

The framework’s intended scope is deliberately narrow. It targets hospitalized medical patients with lung cancer, particularly those admitted to respiratory medicine or medical oncology services. Surgical admissions, brief chemotherapy-infusion stays, intensive care, and end-of-life care require separate adaptation, and patients already receiving therapeutic-dose anticoagulation are excluded from the pharmacological primary-prophylaxis denominator. Components addressing suspected clots support early recognition and escalation, not treatment of established disease. The authors also frame the pharmacist’s role carefully: not as replacing physician decision-making or leading an isolated drug pathway, but as contributing to medication safety, anticoagulant appropriateness review, drug interaction screening, renal and hepatic dose considerations, patient counseling, and continuity of care, a framing that explicitly avoids overstating pharmacist-led effectiveness, which was not tested.

The limitations are candidly acknowledged. The panel of 15 experts came from a single tertiary hospital, with shared departmental affiliations and overlap between interview participants and panelists potentially reinforcing local perspectives. Patients and family caregivers were not represented. The framework has not been prospectively implemented or externally validated, so the study cannot demonstrate reductions in VTE incidence, bleeding, mortality, or costs. Ratings were anonymous to panelists but identifiable to the research team, no structured feedback was provided between rounds, and face-to-face administration may have introduced social-desirability bias. The absence of a standardized quality-appraisal tool for the literature review also introduced subjectivity into item generation. The authors position the framework as a foundation for hospital-based quality improvement and future implementation research, to be tested for feasibility, acceptability, fidelity, adoption, sustainability, and transferability across different hospital levels, with comparative or before-and-after designs needed to determine whether it genuinely improves service delivery and patient outcomes.

Subject of Research: Development of a consensus-based multidisciplinary framework for venous thromboembolism prevention in hospitalized lung cancer patients

Article Title: Development of a multidisciplinary service-delivery framework for venous thromboembolism prevention in hospitalized medical patients with lung cancer: a mixed-methods modified Delphi study

Article References: Meng, Y., Zhong, S., Ding, Y., Xu, P., Wang, Y., Zhang, J., & Wu, Q. (2026). Development of a multidisciplinary service-delivery framework for venous thromboembolism prevention in hospitalized medical patients with lung cancer: a mixed-methods modified Delphi study. Frontiers in Cardiovascular Medicine, 13, Article 1935436. https://doi.org/10.3389/fcvm.2026.1935436

Image Credits: AI Generated

DOI: 10.3389/fcvm.2026.1935436

Keywords: venous thromboembolism, lung cancer, Delphi study, hospital quality improvement, thromboprophylaxis, multidisciplinary care, implementation science, patient safety, clinical pharmacy, risk assessment, quality indicators, Donabedian model