What challenges did AstraZeneca face before AI implementation?

Adverse drug reaction reporting requires gathering and analyzing large volumes of medical literature and producing reports that meet strict NMPA regulatory formatting. The process is time-consuming for researchers.

What was the impact of the AI solution?

Accuracy increased from 90% to 95%. Efficiency in creating adverse event reports increased by 300%.

Source

AlibabaEvidence: Medium50/100

AstraZeneca builds adverse event summary system with Qwen and Dedicated Model Studio

Use case typeMedical writing automationUpdated Jan 1, 2024

AstraZeneca China built an adverse event summary system using Tongyi Qianwen (Qwen) on Alibaba Cloud Dedicated Model Studio. The system ingests medical literature and drafting inputs, generates adverse-event reports to regulatory standards, and requires only human review and minimal edits before sign-off. The implementation uses a private deployment for stronger data/system segregation and privacy protection.

Organization: AstraZeneca
Industry: Pharma
Location: China
Published: January 2024

Reported outcomes

+300%

report creation efficiencyProductivity & throughput

90-95%report accuracy

Strategic outcomes

Risk & complianceAccelerated NMPA-compliant adverse event reportingSpeed & agilityHuman review with minimal editsInnovation & cultureFirst-of-its-kind industry solution

Catalog median for productivity & throughput deployments: +40% across 108 reported metrics. Compare benchmarks →

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Report accuracy: 90-95% increase

Alibaba Cloud Customer StoryJan 1, 2024Customer storyExplicit claimMedium evidence strength

accuracy of the adverse event reports has increased from 90% to 95%

Normalized claim

Report creation efficiency: 300% increase

Alibaba Cloud Customer StoryJan 1, 2024Customer storyExplicit claimMedium evidence strength

efficiency in creating the adverse event reports has increased by 300%

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: AstraZeneca
Provider: Alibaba
Maturity: Unknown
Linked source: Alibaba Cloud Customer Story

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 2 of 2

1Medical writing automation
2Document processing automation

Adverse drug reaction reporting requires gathering and analyzing large volumes of medical literature and producing reports that meet strict NMPA regulatory formatting.
The process is time-consuming for researchers.

Built an adverse event summary system based on Qwen LLM and Dedicated Model Studio.
Integrated AstraZeneca private knowledge bases and publicly available medical and pharmacological knowledge.
Used private deployment on Alibaba Cloud to streamline development and administration while safeguarding data privacy.

Technologies

Qwen Dedicated Model Studio Alibaba Cloud

Accuracy increased from 90% to 95%.
Efficiency in creating adverse event reports increased by 300%.

Architecture

Private cloud deployment on Alibaba Cloud with Tongyi Qianwen (Qwen) LLM and Dedicated Model Studio, integrating AstraZeneca private knowledge bases and public medical/pharmacological knowledge to generate draft adverse-event reports for human review.

Sources & evidence1

Evidence: Medium50/100Evidence strength