AI-Powered VTE Prevention: Insights from an Expert at China-Japan Friendship Hospital
Many nurses have encountered such scenarios: some orthopedic patients undergo successful surgeries, yet suddenly die a few days postoperatively upon getting out of bed. It turns out that these patients developed venous thromboembolism (VTE) after orthopedic surgery, and died from pulmonary embolism caused by thrombus detachment. Once this occurs, there are no specific effective treatments available.
“No one wants to see venous thromboembolism (VTE) occur, yet many physicians have long regarded VTE prevention and management as a ‘thankless task.’”
Venous thromboembolism (VTE) is not a disease specific to any single medical specialty; patients who are postoperative, have suffered trauma, have advanced-stage cancer, are in a coma, or are on long-term bed rest are all at risk of developing VTE. However, assessing a patient’s risk of developing VTE is not straightforward. In addition to relying on physicians’ clinical experience, more authoritative approaches involve using individualized VTE risk assessment models, such as the Caprini Risk Assessment Model and the Padua Prediction Score, to analyze and evaluate patients.
This is precisely where the problem lies. “As a specialist, I have an endless stream of patients to see each day; how could I possibly find the time to conduct VTE risk assessments? Furthermore, the data required for risk assessment is scattered across various departments. Since VTE management is not my area of expertise, it is difficult to perform effective VTE assessments even if I were inclined to do so.”
Such scenarios are not only prevalent in primary care hospitals but also frequently occur in tertiary Grade A hospitals, which boast high-quality resources. Fortunately, as information systems become increasingly unified, clinical pathways are being standardized, and new technologies such as artificial intelligence are continuously adopted, the vast potential that technology brings to the healthcare industry is gradually transitioning from virtual concepts to reality.
To date, many top-tier tertiary hospitals in China have implemented AI-assisted intelligent quality control systems for venous thromboembolism (VTE). By automatically capturing patient data, providing intelligent alerts, and standardizing physicians’ quality control workflows, these systems are gradually transforming the current landscape characterized by inadequate VTE risk assessment and frequent missed or misdiagnoses of VTE.
Recently, VCBeat had the privilege of engaging in an in-depth dialogue with Dr. Xia Lei from the Medical Affairs Department of China-Japan Friendship Hospital, focusing on “VTE prevention and treatment.” The discussion aimed to explore how AI-assisted VTE quality control systems function within hospital settings. Currently, China-Japan Friendship Hospital utilizes the VTE quality control system provided by Beijing Huimei Cloud Technology Co., Ltd.
Since thrombosis presents no obvious symptoms in its early stages and the clinical manifestations after formation are atypical, primarily characterized by swelling and pain, patients often overlook the impact of venous thromboembolism (VTE) on their health, while physicians are prone to misdiagnosis and missed diagnosis.
However, conducting VTE risk assessment and bleeding risk assessment for every patient would undoubtedly add to the burden of physicians who are already under heavy work pressure.
“From a physician’s perspective, common thrombosis risk assessment tools such as the Caprini Risk Assessment Model and the Padua Prediction Score are relatively complex in terms of indicators and time-consuming to score for healthcare professionals who are not VTE specialists. Therefore, even when these assessment scales are used, delayed or inaccurate assessments still frequently occur. If hospitals aim to improve the effectiveness of VTE risk control, only by introducing VTE specialists can quality control and management be made more precise and effective,” explained Dr. Xia Lei.
Furthermore, hospitals urgently need to address the waste of VTE data. Over the years, hospital information systems have accumulated a vast amount of VTE-related diagnostic and treatment information. These data represent an invaluable asset for both internal hospital management and clinical research. However, due to the lack of high-quality information systems, these records remain inaccessible and difficult to analyze, effectively languishing in databases. This constitutes a significant loss for hospitals.
“From my perspective, if we were to prioritize the challenges currently facing VTE screening and prevention, the top three most urgent issues are: the subjectivity of healthcare professionals’ assessments, the time-consuming nature of VTE scoring, and the persistent occurrences of misdiagnosis and missed diagnoses. These issues pose a threat to patient safety. Following closely behind are challenges related to data retrieval and the supervision of prevention and treatment measures. However, all of these problems need to be addressed one by one in the future.”
From the above, it appears that the challenges in VTE prevention and control stem from management issues—hospital administrators, constrained by limited time and manpower, must strike a balance between the quality and quantity of VTE prevention efforts, while also making trade-offs between physicians’ core clinical duties and VTE screening and prevention tasks. However, given the current situation in hospitals, the most critical issue may well be the inherent insufficiency of resources available to administrators themselves.
The intervention of artificial intelligence has undoubtedly changed the game. By introducing a third-party force, hospitals can achieve more extensive quality control, faster scoring, and more accurate judgments, while only consuming a small amount of manpower.
These advantages have helped hospitals rapidly increase their VTE assessment rates. Xia Lei stated, “Previously, VTE quality control relied entirely on the voluntary compliance of doctors and nurses; assessments were conducted only when time permitted. Departments with heavy workloads or those that did not prioritize VTE simply did not perform these assessments.”
“Before the implementation of mandatory AI-driven controls, our hospital’s VTE assessment rate declined month by month. After introducing an AI-based quality control system, the previously stable VTE assessment rate began to decline again. It was only after consulting with VTE specialists that I realized many missed diagnoses had previously gone unrecorded in our statistics. As expected, the assessment rate started to rise rapidly a month later. Currently, we are not only able to perform routine quality control but also conduct dynamic assessments, with each patient being evaluated upon every admission and discharge. Our hospital’s dynamic assessment rate has now reached 200%, significantly reducing missed cases and errors. The integration of artificial intelligence has substantially shortened the time physicians spend on VTE assessments.”
To provide a practical example, if post-assessment results indicate that a patient is at moderate-to-high risk for venous thromboembolism (VTE) and has a significant bleeding risk, mechanical prophylaxis is the physician’s preferred option. Before initiating mechanical prophylaxis, the physician must confirm whether deep vein thrombosis (DVT) has already formed in the patient’s lower extremities. At this stage, Huimei’s VTE quality control system prompts the physician to verify whether compression ultrasonography of the lower extremity veins has been performed. If the deep vein ultrasound result is positive, mechanical prophylaxis is contraindicated, and alternative measures must be adopted.
“Overall, during the period since the implementation of the Huimei VTE Quality Control System, the five issues mentioned earlier have been addressed to varying degrees. Notably, significant improvements have been observed in three key areas: subjectivity in scoring, time-consuming assessment scales, and misdiagnosis or missed diagnoses. Currently, our hospital’s VTE risk assessment and bleeding risk assessment rates are approaching the ultimate target of 100%.”
Furthermore, AI-based clinical decision support functionalities can assist healthcare professionals in conducting automated assessments and issuing alerts to prevent the ordering of contraindicated medications based on patient-specific contraindications. By shifting quality control checkpoints upstream through AI-driven quality assurance, this approach addresses the challenge of managing the intrinsic quality of venous thromboembolism (VTE) care provided by physicians.
Meanwhile, Huimei Technology can import the latest clinical practice guidelines for disease diagnosis into the system. If physicians encounter issues while documenting medical records, they can simply log in to the intranet to access various up-to-date literature. Through this approach, Huimei not only helps hospitals improve VTE quality control but also enhances physicians’ professional competence.
While AI support has elevated VTE prevention and control at China-Japan Friendship Hospital to new heights, addressing VTE is not merely a matter of resolving issues within a single hospital setting. Dr. Xia Lei believes that only by combining subjective changes in awareness among doctors and patients with the objective advancements in information technology can VTE prevention and treatment be more effectively addressed.
Therefore, in addition to implementing in-hospital VTE quality control, we should extend our focus to broader out-of-hospital settings, integrating in-hospital and out-of-hospital prevention and treatment, as well as combining proactive measures by physicians with those taken by patients, to thoroughly eliminate the potential harm of VTE to patients.
“Only when patients and their families understand prevention and control measures can patients achieve better adherence. It is akin to asking children to wash their hands at home; if they are not made aware of the numerous bacteria and viruses present on their hands, they will not take this instruction to heart. Therefore, VTE prevention and control should not be confined to the hospital setting. We hope that enterprises can leverage apps or mini-programs to facilitate online patient assessments, enabling patients to gain a deeper understanding of VTE and bleeding risks, thereby enhancing VTE prevention. Meanwhile, I hope that such assessment results can be integrated into existing VTE systems and directly linked to electronic medical records, allowing physicians to promptly evaluate and issue alerts for patients at moderate-to-high risk of VTE,” stated Dr. Xia Lei.
“Although Huimei’s AI-driven VTE quality control system has significantly improved the prevention and treatment landscape for venous thromboembolism (VTE) at China-Japan Friendship Hospital, we still expect further enhancements. Currently, we maintain ongoing communication with Huimei through weekly meetings, during which Huimei proactively addresses the issues raised. Overall, the journey toward effective VTE prevention and treatment remains long, requiring joint efforts from both hospitals and enterprises.”