Qingsong Health Evidence-Based Medical AI Project Wins National Second Prize in 2026

An evidence-based medical AI project jointly submitted by Sansure Biotech and Beijing Qingsong Yikang Information Technology won second prize at China’s 2026 National Intelligent Perception Competition on Medical Simulators and Health Sensors. The result recognizes a concrete primary-care training use case: a respiratory virtual patient system designed to connect simulated clinical encounters with evidence-based reasoning.

A National Recognition for a Primary-Care Training System

China’s National Healthcare Security Administration published the competition results on August 13, 2026. The official award list placed the project “Evidence-Based Medicine Agent–Powered Primary-Care Respiratory Virtual Patient Teaching System: Improving Primary-Care Respiratory Diagnosis and Treatment Capacity” among the second-prize winners.^[1]

Competition: 2026 National Intelligent Perception Competition on Medical Simulators and Health Sensors
Track: Medical simulators
Category: Digital simulation and AI virtual patients
Group: Virtual patient dialogue simulation
Project number: P0297
Joint applicants: Sansure Biotech Co., Ltd. and Beijing Qingsong Yikang Information Technology Co., Ltd.
Award: Second prize

The award ceremony and outcomes exhibition were held in Changsha on August 23. According to CCTV, the competition attracted 196 projects from 28 provincial-level regions and the Hong Kong Special Administrative Region; 25 first prizes, 37 second prizes, and 49 third prizes were awarded after preliminary and final evaluation.^[2]

Beijing Qingsong Yikang is identified by the China Securities Regulatory Commission as a domestic operating entity of Qingsong Health Corporation.^[3] QSEvidence is Qingsong Health’s evidence-based medical AI agent, publicly positioned around medical literature, clinical guidance, source traceability, and reviewable medical workflows.^[4]

What the Competition Actually Evaluated

This was not a contest for polished demonstrations alone. The medical simulator track covered physical simulators, digital simulation systems, AI virtual patients, and embodied nursing robots. Official competition materials required digital systems to be deployable and functionally complete rather than concept-only submissions.^[5]

The Hunan Provincial Healthcare Security Administration described three central judging questions for medical simulators:

Does the product match real medical education or clinical practice scenarios?
Is its medical logic rigorous and integrated with an authentic healthcare workflow?
Does it demonstrate innovation and a realistic path to implementation?

At the final, medical simulator projects drew tasks from a test bank and demonstrated their systems live. Clinical and technical experts then evaluated the product presentation, underlying logic, and answers to questions. The process included archived scores, test data, and audiovisual records for traceability.^[5]

Why a Respiratory Virtual Patient Matters in Primary Care

Respiratory complaints are common, but an apparently simple consultation can require several linked judgments. A learner must gather a focused history, identify red flags, interpret symptoms in context, decide what further information is needed, and distinguish between conditions that can be managed locally and those requiring escalation.

A virtual patient can make those decisions observable. Instead of memorizing an answer, the learner interviews a simulated patient, chooses questions, receives evolving information, explains a differential, and sees how an early assumption changes the next step. When evidence retrieval is added to the debrief, the training can also ask a more important question: What supports this decision?

The second-prize project is therefore notable for connecting three components that are often separated:

Simulation creates a repeatable clinical encounter.
Dialogue exposes the learner’s information-gathering and reasoning process.
Evidence support gives instructors and learners a path back to guidelines and literature.

What the Result Means for QSEvidence

The award belongs to the named joint project and its participating organizations. It should not be interpreted as a blanket certification of every QSEvidence output, nor as permission to use AI without professional review.

It does, however, reinforce the practical direction behind QSEvidence: medical AI becomes more useful when it supports a defined workflow and keeps the evidence path visible. In a training setting, QSEvidence can help educators prepare source-linked case materials, compare relevant guidance, identify points of uncertainty, and create a debrief package that instructors can review before it reaches learners.

That is a more demanding role than generating a plausible answer. The system has to help connect the learner’s question, the simulated case, the evidence retrieved, the interpretation made, and the human reviewer’s final judgment.

From an Award-Winning Concept to Routine Training

Institutions considering a similar model can begin with a narrow, auditable implementation:

Select one high-frequency scenario. Start with a respiratory presentation that has clear learning objectives and escalation boundaries.
Write the evidence map before the dialogue. Define which guideline statements, studies, and local protocols support each decision point.
Test clinical logic, not conversational style. A fluent virtual patient is not useful if its physiology, timeline, or response to learner actions is inconsistent.
Require a structured debrief. Show which questions changed the assessment, which evidence was applicable, and where uncertainty remained.
Keep expert oversight. Clinical educators should approve cases, monitor updates, and review unexpected model behavior.

This approach keeps the technology tied to a measurable educational purpose: improving the quality and consistency of clinical reasoning practice without presenting simulation as a substitute for supervised patient care.

Frequently Asked Questions

Did QSEvidence itself win the second prize?

The official award list names a joint project submitted by Sansure Biotech and Beijing Qingsong Yikang Information Technology. Beijing Qingsong Yikang is a domestic operating entity of Qingsong Health, the group behind QSEvidence. The award should be attributed to the named project and applicants.

What was the project designed to do?

Its official title describes an evidence-based medicine agent–powered virtual patient teaching system intended to strengthen respiratory diagnosis and treatment capacity in primary care.

Why combine an evidence agent with a virtual patient?

The virtual patient provides a repeatable encounter, while the evidence layer can support case construction and debriefing with reviewable sources. The combination helps training move from answer recall toward explainable clinical reasoning.

Can a virtual patient replace bedside teaching?

No. Virtual patients can supplement practice, especially for repeatable scenarios and structured feedback, but they do not replace supervised clinical experience, real patient communication, or educator judgment.

Can the award be treated as proof of clinical effectiveness?

No. A competition award recognizes the evaluated project under the competition’s rules. Clinical or educational effectiveness still requires transparent validation, appropriate outcome measures, local implementation review, and ongoing monitoring.

References

Medical and educational disclaimer: This article reports publicly available competition information and discusses a training workflow. It does not provide medical advice, certify clinical performance, or replace clinician, educator, institutional, regulatory, or patient-safety review.