📌 01 · Healthcare challenges & the value of big data

Global issues Population aging, rising chronic diseases (China: ~114 million diabetics).

Traditional R&D costs increased by 145% from 2004 to 2014.

Shortage of skilled healthcare workers due to demographic shifts.

China-specific dilemmas Supply-demand imbalance: only 2.31 licensed physicians per 1,000 people (2016).

Ranking of China in physician density lies between 25th-30th among OECD nations.

Resource misallocation: tertiary hospitals’ visit CAGR 10.7% vs 1.5% at primary level (2010-2016).

Medical insurance deficit projected: cumulative deficit of 735.3 billion RMB by 2024.

McKinsey estimated that big data in US healthcare could reduce annual medical costs by $300-450 billion.

Big data will upgrade China's medical ecosystem from infrastructure to clinical innovation and preventive care.

📈 02 · Macro tailwinds — policy, funding, tech

Policy push : Since 2015, State Council and NHC issued "Big Data Development Action Plan" and "Healthy China 2030".

Health big data elevated to national strategic level.

Technology Ministry allocated over 1.2 billion RMB for precision medicine and chronic disease projects (2017-2018).

Pilot data industrial parks established in Fujian, Jiangsu, Shandong, Guizhou.

Capital momentum : In Q1 2018, health big data saw 35 investment events.

AI-assisted diagnosis attracted significant capital; 17 financing rounds in 2017 for decision support.

Estimated market size of AI-aided diagnosis to reach 5.586 billion RMB by 2022.

Technology maturity : Global health data growing at 48% annually; over 2300 exabytes by 2020.

Hospital information systems penetration: 70%-80% in tertiary hospitals.

Deep learning and NLP enable extraction from unstructured data (images, reports).

🏥 03 · Big data classification & healthcare value chain

Three data domains : In-hospital data (clinical, behavior), out-of-hospital (wearable, health records), genomics (exome, whole genome).

Scenarios often combine multiple data sources: early warning, precision medication, imaging AI.

Value chain : Upstream data providers (hospitals, cloud storage); Midstream AI/core tech firms (image recognition, NLP); Downstream B2B clients (hospitals, pharma, insurers, government) and eventually consumers.

Geographic concentration: Beijing, Guangzhou, Shanghai due to top-tier hospitals and local government support.

In 2022, about 10% of medical institutions may purchase assisted diagnosis solutions.

⚙️ 04 · Key application scenarios in depth

🔍 Clinical decision support (imaging & general)

Imaging AI for lung nodules, breast cancer, fundus diseases: accuracy over 90%.

General CDSS accuracy ~85%, still improving.

Revenue models: software/hardware solutions, SaaS rental, per-use fee in the future after Class-III medical device certification.

📱 Out-of-hospital & chronic care management

Online consultation volume: 23.8 million consultations in 2017 (Frost & Sullivan).

Ping An Good Doctor’s AI assistant reduces labor costs; paid conversion rate 2.7% in 2017.

Miao Health integrates data from 300+ smart devices, provides B2B2C health management for insurers and corporates.

Lanxinkang offers pharmacy-centric chronic disease SaaS, covering >300 cities.

🏥 Smart hospitals & regional information platforms

Yuanqi technology builds hospital external connection platforms, accumulating registration/payment big data.

Kingdee Medical’s regional EMR sharing platform supports tiered care in Foshan.

National health big data groups (China Electronics Corp) lead data aggregation and governance.

🏢 05 · Leading companies & case studies

✨ Ping An Good Doctor (01833.HK)

200M+ registered users, raised 1.1 billion USD in IPO (2018).

"Modern Huatuo Plan" standardizes TCM with AI; AI assistant improves triage efficiency.

"One-minute clinic" integrates smart triage, AI prescription, remote consultation.

🧠 InferVISION (推想科技)

Deep learning for lung, stroke, liver cancer imaging.

Deployed in nearly 100 top-tier hospitals (PUMCH, Tongji, Changzheng).

Daily AI lung screening ~10,000 cases; click rate in primary hospitals up to 100%.

🔬 DeepCare (羽医甘蓝)

Cervical cancer, breast cancer, oral pathology AI with automatic lesion annotation.

Cooperates with 19 tertiary hospitals; launched skin defect detection device (2018).

❤️ Miao Health & Lanxinkang

Miao: AI-driven health behavior intervention, cooperation with China Unicom, Huawei, Sun Insurance.

Lanxinkang: SaaS for pharmacy chronic care, improved medication adherence; published 10+ real-world studies.

📡 Internet giants (Alibaba, Tencent, Baidu)

Alibaba Health "Doctor You" & ET Medical Brain; Tencent "Miying" AI for cancer screening, investment in Atomwise.

Baidu shifted to AI drug discovery.

🚀 06 · Challenges & emerging trends

⚠️ Present challenges

Regulatory lag: CFDA Class-III certification pathway still under development for AI clinical applications.

High data labeling costs (physician labor, up to 30 min per image).

Data silos persist across institutions.

Weak C-side payment willingness; monetization largely relies on B2B.

📌 Future trajectories

B2B first, B2C later : Hospitals, pharma, and insurers are early adopters; national teams leading infrastructure.

Mid-term: consumer-paid preventive care, genetic testing subscription, per-visit AI diagnosis fee.

AI bridges the gap from raw data to actionable insights: NLP + computer vision unlock unstructured data.

Ecosystem collaboration becomes the winner: single companies cannot cover full patient journey.

📎 Appendix · China healthcare fundamentals

• Aging population: 158 million aged 65+ in 2017 (11.4%), driving medical demand.
• Physician shortage: China ranks 25-30 globally in physicians per capita; top-tier doctors heavily concentrated.
• Medical insurance stress: By 2024, cumulative deficit is predicted to reach 735 billion RMB.
• Commercial health insurance and big-data based actuarial science will expand rapidly.