Study design and setting

iWAVE is a prospective, multicentre, observational, population-based cohort study initiated at West China Hospital, West China Tianfu Hospital, and Shang Jin Hospital of West China Hospital. The project will progressively expand to include the 15 hospitals within the West China Hospital Medical Consortium and their affiliated community hospitals for ongoing participant recruitment (figure 1). The National Clinical Medical Research Center for Geriatric Diseases and the Heart Valve Disease Research Center at Sichuan University’s West China Hospital will manage the study’s design and implementation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Distribution of West China Hospital and units of the 15 participating medical consortiums.

Study population

The categorisation of participants on the basis of the source population is detailed below. (1) Community-based natural population cohort: Eligible participants are permanent residents within the jurisdiction of the West China Hospital Medical Consortium who agree to undergo transthoracic echocardiography at community hospitals. Sonographers conducting the procedures will receive standardised training from the Departments of Ultrasound and Cardiology at West China Hospital. The inclusion criterion is that the participants must be 35 years of age or older. The exclusion criteria include an inability to comprehend the questionnaire or refusal to participate in the study. (2) Inpatient and outpatient follow-up cohort: This group consists of participants from the cardiovascular surgery departments of West China Hospital and 15 associated medical consortiums. The eligibility criteria include patients initially evaluated with moderate or more severe VHD via echocardiography, being 35 years of age or older, and having the capacity to understand the questionnaire. The exclusion criteria include impaired consciousness or death that prevents the completion of the questionnaire, a life expectancy of less than 6 months, acute aortic dissection and refusal to participate in the study.

Study procedure

The data elements and definitions for the iWAVE Cohort Study form were meticulously designed, drawing inspiration from the current STS ACSD database template and adapting it to the specific needs of this study. Figures 2 and 3 provide detailed information on the study process.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Recruitment process of the Integrated Whole-Life Cycle Accuracy Valvular Heart Disease Epidemiology (iWAVE) Cohort Study. EDC, electronic data capture.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Summary of clinical data collection.ECG:Electrocardiogram; 4D-CT: 4-Dimensional Computed Tomography; 6MWT: 6-minute Walking Test; PFT: Pulmonary Function Test; KCCQ: The Kansas City Cardiomyopathy Qusetionnaire; EQ-5D: EuroQol Five Dimension Scale; EQ-VAS: EuroQol Visual Analogue Scale; PSQI: Pittsburg Sleep Quality Index; HADS: Hospital Anxiety and Depression Scale. Mandatory; Available.

To streamline data collection and organisation, we developed a Chinese electronic data capture (EDC) system. This system requires network connectivity via a computer or handheld mobile device, with each hospital site authorised to access the EDC system via an assigned account. The system accommodates various unit levels, including community hospitals within medical unions, medical union units and the central hospitals of West China Hospital, Sichuan University. To address the unique characteristics of patient diagnosis and treatment across different levels in the western region of China, specific electronic forms within the EDC system will be designated for data collection at each unit level. To comprehensively delineate the characteristics of patients with VHD, community hospitals affiliated with the medical consortium will primarily gather community-based health examination data from patients. This information will facilitate patient screening and categorisation, guiding regular follow-ups within the community on the basis of disease severity. In contrast, the medical consortium and West China Hospital, Sichuan University, will primarily collect diagnostic and treatment data during hospital visits. The integrated care management system will include patient follow-up components with scheduled follow-ups at 1 month, 3 months, 6 months and 1 year postdischarge, followed by regular annual outpatient follow-ups, addressing instances involving adverse events that require medical attention.

Trained VHD data collectors will conduct daily eligibility screenings for patients in outpatient clinics and hospital wards. On identifying eligible patients, the EDC system will automatically compile their essential information. This includes the patient registration number, which serves as a unique identifier within the treating hospital; the outpatient visit number for the current visit; the inpatient case number corresponding to the ongoing hospitalisation; the patient’s ID card number as a distinctive patient identifier; and the phone number as the primary contact method. Within the EDC system, a new account is established for each eligible patient, with these details recorded as essential information. When new VHD patient data are entered into the EDC system, a unique index identifier is automatically generated for the patient. The system then correlates the patient’s medical records within the medical consortium units on the basis of the patient’s ID card number. It autonomously collects patient treatment data using the registration number, visit number or case number, thereby constructing comprehensive medical records within the medical consortium.

Continuous follow-up of patients is a crucial aspect of this study, addressing the scarcity of long-term follow-up data for patients with VHD both domestically and internationally. Trained physicians will conduct ongoing follow-ups and gather patient-related information through face‒to‒face interviews via touch screen questionnaires or paper records. Full-time data collectors execute the follow-up data collection, meticulously verifying and standardising questionnaire responses before entering them into the EDC system to ensure accuracy and consistency. The EDC system dispatches follow-up reminders to data collectors 1 week before each scheduled follow-up (at 1 month, 3 months, 6 months and 1 year postdischarge, among others). We will make three attempts to contact patients on different dates and times, depending on the availability of the contact person, as indicated by phone records. All subsequent calls will be recorded for verification, and regular quality control assessments will be conducted on the follow-up results. In cases where patients cannot be contacted, we plan to undertake additional steps to obtain long-term health-related outcomes on the basis of personal information. This involves correlating the patient’s ID card number with data from the Chinese Center for Disease Control and Prevention’s national disease surveillance system, the cause of death surveillance network report database, the inpatient data identifier from the community hospital in the patient’s residential area, and regional death registration records. If these follow-up methods are unsuccessful, the case will be categorised as a loss to follow-up. Notably, participants who are unable to complete follow-up procedures will still be contacted for subsequent follow-ups unless they actively request the cessation of communication. The data collected during follow-up can be linked with patient records in the EDC system using the patient’s ID card number as the matching identifier. Ultimately, the patient’s timeline for medical and outpatient follow-ups can be organised according to treatment and follow-up dates.

To ensure the scientific rigour of the study, a consulting committee comprising experts in cardiac surgery, interventional cardiology, advanced imaging cardiology, echocardiography, heart failure, critical care, cardiac anaesthesia, cardiorespiratory physiotherapy, specialist disease management, statistics, computer engineers and other specialised fields was established. Additionally, a project management and coordination centre was created to oversee and support the iWAVE operation, conduct data quality control, and facilitate regular online and offline data quality control meetings. An independent supervisory committee was also appointed to oversee the entire research process.

Outcomes

Data collection

The following data will be collected for all participants in both the inpatient and outpatient follow-up cohorts, as well as the community-based natural population cohort:

1. Basic demographic information and peripheral blood samples.

2. Full life cycle medical imaging data.

3. Long-term functional follow-up and patient self-reported outcome questionnaires.

1. Inpatient and outpatient follow-up cohorts:

i. Peripheral venous blood parameters, including cardiac biomarkers, biochemical indices, coagulation, routine blood cell analyses, endocrine hormone measurements, glycosylated haemoglobin levels, inflammatory factors, and samples for cell-free DNA analysis, miRNA analysis, cytokines, telomere length, metabolomics, proteomics, transcriptomics, and genomics.

ii. Routine biochemical indicators of urine and stool and samples for precision medicine tests, such as microbial sequencing.

iii. Myocardial and/or valvular tissue samples.

iv. Blood pressure, 12-lead ECG and peripheral oxygen saturation.

v. Echocardiogram and Doppler measurements and cardiac 4-dimensional CT (4D-CT) measurements.

vi. Electronic health record information, including comorbidity information, medication use, surgical risk assessment, surgical operative records, anaesthesia monitoring data, critical care records and in-hospital adverse outcome events.

vii. Valve clinical study endpoint events tracked throughout the patient’s lifetime.

viii. Functional capacity measures, including the 6 min walk test (6MWT), pulmonary function test (PFT) and New York Heart Association (NYHA) classification.

ix. Patient self-reported outcome questionnaires: The Kansas City Cardiomyopathy Questionnaire-12 (KCCQ-12) Questionnaire; EuroQol Five Dimension Scale (EQ-5D) and EuroQol Visual Analogue Scale Questionnaires; lower extremity oedema level; Pittsburgh Sleep Quality Index (PSQI); General Anxiety Disorder Scale-7 (GAD-7); Patient Health Questionnaire-9 (PHQ-9).

x. Home wearable devices to capture continuous physiological parameters.

1. Community-based natural population cohorts:

i. Peripheral venous blood analysis, routine biochemical indicators in urine and stool, and sample collection for precision medical testing.

ii. Blood pressure, 12-lead ECG and peripheral oxygen saturation.

iii. Echocardiograms and Doppler measurements.

iv. Electronic health record information, including comorbidity information and medication utilisation.

v. Major adverse cardiovascular events tracked throughout the patient’s lifetime.

vi. Functional capacity measures, including the 6MWT, PFT and NYHA classification.

vii. Patient self-reported outcome questionnaires.

Sample size

On the basis of the literature review and the research team’s preliminary retrospective data, the 5-year average survival rate for patients with VHD managed via conventional protocols and who are undergoing surgical interventions is approximately 60%.10 We anticipate a 30% improvement in this survival rate with the implementation of a novel integrated care management protocol for patients with VHD. With a superiority ratio of 1.2, a test level of 0.025, a test efficacy of 0.90 and a de-escalation rate of 20%, at least 1889 VHD inpatients and outpatients need to be enrolled. Furthermore, this study aims to investigate the prevalence of VHD within the broader community. According to a survey, the weighted prevalence of VHD among the Chinese population aged 35 years and older is 3.8%.1 The likelihood of the sample rate obtained during the formal survey differing from the known prevalence by no more than 1% does not exceed 0.05. Considering a deleterious rate of 20%, a minimum of 7270 patients from the community-based population is required.

The West China Hospital of Sichuan University is one of the largest tertiary hospitals in western China, serving a population that constitutes approximately 6% of the country’s total population within its 15 medical consortiums and associated subcommunity hospitals.39 Therefore, we plan to enrol at least 10 000 patients over a 5-year period (1 October 2023 to 1 October 2028) to achieve the objectives of this prospective cohort study.

Data management and monitoring

The information stored in the iWAVE Registry Database is unalterable and can only be accessed for query analysis through database views. The EDC system provides standardised forms for VHD follow-ups and is available in both mobile and personal computer versions. Trained data collectors will be authorised to conduct follow-ups and collect data exclusively within the hospital’s internal network environment, using tablets or physician workstations. Data integration will occur through a dedicated network channel in the medical cloud to ensure data security.

All collected data will undergo daily synchronisation, where information from the hospital’s electronic health record system will be automatically integrated with the EDC system. This synchronisation will retrieve comprehensive data about participants' medical management, laboratory test results and, if applicable, surgical procedures throughout the research period. To establish a connection between health records in a fully covered participant management database for long-term medical or mortality records 1 year after recruitment and beyond, an annual linkage will be implemented (figure 2).

The EDC system data collection form template was developed by the Heart Valve Team at West China Hospital, Sichuan University. All units within the medical consortium, including community hospitals, will adhere to this standardised template for data collection. Before data collection begins, researchers and data collectors will undergo comprehensive training covering key elements, definitions and the data input system. The Project Management Coordination Center will conduct regular meetings to provide feedback on data quality control reports and implement retraining as needed. Additionally, there will be a strong emphasis on maintaining the stability of the onsite team members. When there are changes in onsite researchers, coordinators or data collectors, prompt training for new staff becomes imperative.

Participating sites will ensure the reporting of all eligible cases to iWAVE, with the specific methods outlined below. For sites using electronic medical records, an electronic query algorithm will be employed to identify patients with VHD by searching for data fields that are most likely to indicate relevant events among all patient records involved in cardiovascular surgery. These events may include conditions such as ‘rheumatic heart disease’, ‘degenerative heart disease’, ‘valvular stenosis’ and ‘valvular regurgitation’. Coordinators will subsequently evaluate compliance with the inclusion criteria on the basis of the electronic medical records. In cases where electronic medical records are not used or lack statistical functionality, such as when patients present external examination reports or hospital records, the coordinators will manually review the patients’ external paper records. The accuracy and completeness of the reported data will be ensured through a multifaceted approach. First, the EDC system incorporates data input checks to mitigate errors and enhance the precision of data collection. An algorithmic quality check system automatically prompts notifications for researchers to rectify or review data deviations from predefined rules. Second, quality inspectors will meticulously examine each variable on the basis of medical records within 7 days of data submission. If incorrect, suspicious or incomplete data are identified during ongoing audits, researchers reporting on patients with VHD must correct or complete the records and clarify any missing or dubious data elements. The inspectors will summarise the monthly progress of data collection and quality measures and engage in regular discussions with the principal investigator. Third, statisticians will routinely evaluate and screen the collected data, including checks for logic errors and outliers. Fourth, the project management centre will conduct random testing on 10%–50% of the data weekly to examine extreme values or missing entries. Fifth, the project management centre will deploy a coordinator to hospital wards and outpatient clinics every 3–6 months for onsite checks, ensuring the continuity, completeness, and accuracy of the reported data. Finally, the experts will randomly review more than 20% of the audio recordings from the entire interview process to ensure the quality of the baseline and subsequent data. In cases of disagreement, the data collectors and quality inspectors will collaboratively reassess the data (figure 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Data management and quality control. EDC, electronic data capture.

Data security

The EDC system allows access exclusively within the confines of the hospital’s internal network, with access permissions tailored to the specific responsibilities of individual researchers. Each researcher is assigned a unique account, and user authentication is achieved through secure passwords, thereby reducing the risk of unauthorised access to the EDC system. In accordance with ethical standards, the EDC system employs anonymisation and deidentification processes for reported data. All data transfers are securely conducted within encrypted tunnels and stored in the database of West China Hospital, Sichuan University Information Center. To minimise the risk of data loss, routine backups are systematically performed for all storage.

Statistical analysis

Data will be presented as percentages for categorical variables and as the means±SDs or medians (25th to 75th percentiles) for continuous variables. The distribution of continuous variables will be examined via the Kolmogorov‒Smirnov test. Comparisons will be made using Student’s t-test or the Mann‒Whitney U test for continuous variables and the χ² test or Fisher’s exact test for categorical variables, as appropriate. Linear mixed-effects models will be employed to assess changes in clinical parameters (eg, echocardiographic measurements, indicators of functional capacity) from the time of surgical intervention to long-term follow-up.

Survival analyses will be performed using the Kaplan‒Meier method to report annual event rates. Model survival functions will be estimated using risk ratios associated with predetermined key covariates adjusted for age, sex and other potential confounders, including comorbidities and relevant echocardiographic characteristics. The influence of prognostic factors on patients with VHD will be explored through multivariate Cox proportional hazards regression modelling, with the degree of influence quantified via HRs with 95% CIs. The selection of variables for the model will be based on their clinical importance and will include process variables such as VHD-specific characteristics, preoperative assessments, surgery-related information and intensive care management. Independent predictors will be selected using Stata’s Lasso technique with cross-validation.

Binary logistic regression, which is based on L1 and L2 regularisation, will be applied with the occurrence of the outcome event as the dependent variable. The coefficients from the logistic regression model rank the importance of each parameter, constructing a feature screening model on the basis of the importance of the parameters and the model’s predictive power. The best subset of features will be determined by iterating through the extracted features for selection until stability in the feature set is achieved. Fivefold cross-validation will be used to generate the training and validation sets, where the training set builds the model and the validation set confirms the model’s capabilities. We will experiment with the selected features using various machine learning models, including logistic regression, support vector machine, linear discriminant analysis, k-nearest neighbours and random forest classifiers, and employ fivefold cross-validation for training. The predictive power of the models will be assessed by calculating the area under the receiver operating characteristics curve. Statistical analyses will be conducted using R V.3.6.2. A two-tailed value of p<0.05 will be considered statistically significant.

Patient and public involvement

Patients and the public are not involved in the design, recruitment or conduct of this study. While we do not plan to inform every individual participant of the study results, we will disseminate the findings to the public through community outreach and public engagement initiatives.

Ethics and dissemination

This study was approved by the Biomedical Research Ethics Committee of West China Hospital of Sichuan University (No. 20232422). All research organisations contributing data for the iWAVE Study will comply with this ethical approval, and any modifications to the study protocol will be submitted for further ethical review. All participants will be required to provide written informed consent, and patient data will be stored in deidentified form at the Information Center of West China Hospital of Sichuan University. All protocols will be conducted in compliance with the Declaration of Helsinki. Access to the study data will be restricted to researchers and ethics committees associated with the study.

The results will be disseminated in various formats, including conference abstracts, posters, presentations and peer-reviewed scientific journals. All of the results generated in this study will be cohort data, ensuring that participants cannot be individually identified. Additionally, the findings will be reported to local governments and health and wellness committees to inform policy development, as well as to medical research centres and units that participate in and support the cohort study.

Discussion

To the best of our knowledge, this is the first cohort study conducted across multiple medical consortiums led by a regional centre hospital in Western China. It focuses on the diagnosis, treatment and prognostic outcomes in a population with VHD in the region, providing comprehensive data on risk factors, diagnostic and therapeutic management, physical status, and whole-life cycle follow-up experiences.

Although national databases for cardiac surgery have long been established internationally, there are notable differences in population and disease characteristics, as well as healthcare public service systems, among China, Europe and the USA.14 26 Therefore, the iWAVE cohort database accounts for the unique population characteristics of western China and focuses on screening variables related to the diagnosis, treatment and prognosis of VHD.20 The iWAVE database will enable continuous monitoring of the entire lifespan of patients with VHD within this population. A series of analysis reports based on the iWAVE cohort study will provide a comprehensive overview of the incidence, demographics, care pathways and whole-life cycle follow-up of VHD in the population of western China. This report will identify risk-modifying factors that influence outcomes, establish benchmarks among cardiac centres, and propose system-wide strategies aimed at improving the survival of patients with VHD. Additionally, the database is promising for assisting both the public and the government in prioritising primary prevention and lifelong follow-up of patients with VHD. Successful outcomes may motivate other cardiovascular medical centres across the country to participate, potentially establishing a national platform for monitoring cardiovascular surgical care in western China, facilitating research projects, establishing benchmarking standards and identifying areas for potential quality improvement.

This study has several limitations. First, the medical centres involved are confined to western China, resulting in a cohort that represents only the VHD population in that region and fails to capture the broader characteristics of VHD across China. However, as the study progresses, we aim to broaden the inclusion criteria to encompass more large medical centres in non-western China, thereby diversifying the geographical distribution of the study population over time. Second, the current biomarkers lack specificity in predicting VHD disease progression, and with the ongoing advancement of transcatheter valve interventions, obtaining patient biospecimens has become increasingly challenging. Therefore, we plan to collect peripheral venous blood samples and urine and stool samples from a community-based natural population cohort, as well as samples of valvular and/or myocardial tissue from patients who undergo median open-heart surgery at the medical centres involved in this study, to explore the microscopic mechanisms of VHD disease. Third, the design of the cohort study limits our ability to implement various medical interventions for conducting long-term prognostic correlation analyses. Nonetheless, incorporating a multicentre cohort population could provide a sufficiently large sample size for propensity score-matched analyses aimed at exploring the relationships between target variables and outcome events. Finally, significant time and resources were dedicated during the prestudy phase to ensure the accuracy and completeness of the data. In future sustainability studies, the iWAVE cohort study is committed to streamlining the data reporting process and automating data collection.

Data availability statement

We invite national and international collaborations for projects concerning various aspects, including but not limited to disease screening strategies, diagnostic and therapeutic decisions, and long-term functional trajectories. Researchers are encouraged to apply for data access by contacting the chief investigator (YG) at drguoyq@wchscu.cn, providing details of the study objectives and proposals.