How YaoLing MedTech Combines AI with CRO to Reduce Data Processing Costs to One-Tenth

Contract Research Organizations (CROs) are the product of specialization, segmentation, and risk diversification in the pharmaceutical industry, primarily aimed at enhancing R&D efficiency. However, numerous inefficient links within the CRO industrial chain continue to hinder industry development. To accelerate the new drug development process, AI has been introduced to address these inefficiencies, with drug discovery serving as its primary battlefield. This article introduces Yaoling Medical Technology, a CRO company that applies AI to transform data collection business processes.

As a technology-driven CRO organization, Yaoling Medical Technology applies artificial intelligence to the field of clinical research, leveraging cutting-edge IT and AI technologies to enhance the efficiency of data collection in clinical studies and significantly reduce the costs associated with data collection and processing.

Amid the booming development of AI, how to deeply integrate AI with vertical niche sectors—transforming traditional workflows while helping professionals improve work efficiency—has become a hot topic in the industry.

In response, VCBeat conducted an exclusive interview with Mr. Hu Qitong, CTO of Beijing Yaoling Medical Technology, to reveal how this healthcare technology company has leveraged AI to reduce overall costs to one-tenth of previous levels while simultaneously enhancing the quality of research evidence.

In clinical trials, Phase IV trials encompass the post-marketing application research stage for new drugs. Their purpose is to evaluate the drug’s efficacy and adverse reactions under conditions of widespread use, assess the benefit-risk profile in general or special populations, and optimize dosing regimens. Yaoling Medical Technology currently focuses primarily on Phase IV clinical trials and post-marketing studies.

The core business activities of clinical trial CROs can be categorized into: clinical trial technical services, clinical trial data management services, clinical trial statistical analysis services, and regulatory submission-related services. Among these, clinical trial technical services constitute the primary business segment, while clinical trial statistical analysis services involve the highest level of technical expertise and demonstrate stronger profitability.

Clinical trials involve multiple stakeholders, including hospitals, pharmaceutical companies, and patients. This is not the first time AI has partnered with the field of new drug development; it has already been applied in compound discovery and targeted drug discovery. In more niche areas, such as foreign patient communities, AI is also being used to match patients for clinical trials. Regardless of how pharmaceutical companies collaborate with AI, the goal is to improve the overall efficiency of new drug development and reduce its costs.

On average, new drug development takes 10 years, much of which is wasted due to the selection of inefficient sites or study designs. In a previous interview with Ross, President of Medidata, VCBeat reported that, according to their research, nearly all of the 30 clinical trials examined had data quality issues. Most trials showed discrepancies in adverse event reporting, 90% exhibited data inconsistencies, and 30% indicated potential operational errors.

Traditional processes not only incur substantial time costs but may also directly lead to the failure of an entire project. Regulatory authorities are increasing personnel investment and training, aiming to accelerate approval processes while aligning approval standards with international norms. Poor data quality at Contract Research Organizations (CROs) can not only prolong regulatory review timelines but also significantly increase the risk of approval failure.

YaoLing Medical Technology has currently developed three major systems to address issues related to data collection, processing, and standardization: the Verify system, the RDCS system, and the Million Data system. The Verify system primarily addresses low-cost data collection; it currently enables tasks to be completed using only one-tenth of the original time and cost.

Hu Qitong stated to VCBeat, “Numerous issues frequently arise during clinical data collection, such as invalid and incomplete data. The RDCS system not only performs data cleaning but also standardizes the data.”

Currently, Yaoling Medical primarily serves four categories of clients: domestic pharmaceutical companies, contract research organizations (CROs), and contract sales organizations (CSOs). In addition, Yaoling Medical Technology also serves various pharmaceutical associations to explore real-world studies.

Throughout the entire process of data collection, structuring, and standardization, AI has played a pivotal role at every stage.

In terms of data collection, Yaoling Medical Technology’s data primarily comes from two sources: one is integration with hospital HIS systems, and the other is through OCR recognition of original medical data files. The aspect where AI can provide support lies in the application of NLP (Natural Language Processing) technology.

Hu Qitong explained, “Although OCR technology is a relatively mature module, its word error rate is approximately 20%–30%. For clinical trials, such an error rate can significantly impact subsequent data structuring or EDC form entry. Therefore, we have added a medical language post-processing module after OCR recognition. This module leverages specialized vertical medical language models to correct misrecognized words, thereby improving recognition accuracy.”

In terms of standardization, the previously fragmented and isolated approaches to data collection, representation, and recording have rendered a large volume of data as “dead data.” Yaoling Medical Technology has established an exclusively self-developed knowledge graph system within its RDCS platform, capable of unifying diverse data representations into a single standardized format.

Hu Qitong provided a simple example to illustrate this: “For instance, an upper respiratory tract infection might be referred to as ‘cold with fever’ in some contexts, and terminology can vary across hospitals. Previously, such cases relied entirely on manual annotation; now, with standardized systems like RDCS, they can be unified under a single concept.”

After data standardization, the data can be reorganized and processed. Value is unlocked only through matching between datasets.

The Million Data system further processes data using statistical models, such as the t-test, Cox regression model, and Kaplan-Meier (KM) survival curves. By providing heuristic guidance through these models, the system assists various departments within pharmaceutical companies and clinical researchers in conducting data analysis and statistics, thereby facilitating intelligent decision-making.

In addition to improving efficiency through digitalization, Yaoling Medical Technology also employs structured approaches to compress the time required for entire clinical trials.

In terms of structure, Yaoling Medical Technology employs machine reading comprehension to automatically identify content and populate it into the corresponding Electronic Data Capture (EDC) system. The standardization process is facilitated through knowledge graphs.

Hu Qitong explained to VCBeat, “Some terms require mapping hypernyms to hyponyms. For instance, ‘bronchitis’ may share no lexical overlap with ‘respiratory diseases,’ yet it is essential for machines to accurately comprehend this relationship at a cognitive level.”

There is a common saying in the CRO industry: “The ceiling of the industry’s market size is determined by pharmaceutical companies’ R&D expenditure.”

Currently, China has emerged as the world’s second-largest pharmaceutical market and is the fastest-growing economy in the global drug sector. Against the backdrop of favorable policy initiatives, both multinational corporations and domestic enterprises are placing greater emphasis on R&D investment.

For pharmaceutical companies, as the difficulty of new drug development increases and international pharmaceutical enterprises experience declining profits ahead of patent cliffs, these global players are increasingly integrating CROs into their R&D processes to control costs, shorten development cycles, and mitigate research risks.

For domestic pharmaceutical companies, the drug review reforms implemented by the NMPA and the National Healthcare Security Administration—such as the Two-Invoice System, volume-based procurement, priority review and approval for innovative drugs, and consistency evaluation for generic drugs—will bring unprecedented development opportunities to China’s CRO industry. Additionally, while Chinese pharmaceutical companies have historically prioritized sales over R&D in their investment strategies, this landscape is now shifting.

According to data on R&D investment from a global sample of 2,500 pharmaceutical companies during the 2015–2016 fiscal year, compiled by the European Commission, Chinese pharmaceutical companies recorded the fastest growth in R&D spending at 27.5% compared with 2015, far exceeding the global annual growth rate of 9.8%.

Amid the unprecedented opportunities for growth in the CRO sector, enhancing data collection quality has become an inevitable trend. Many CRO companies are undergoing transformations in this area, but they predominantly opt for external IT firms or adopt development solutions provided by cloud computing companies.

To better address the pain points of multiple stakeholders, IT companies must first be capable of providing comprehensive solutions; however, integrating into hospital systems is a challenging process. Additionally, they may lack cost advantages. In contrast, Yaoling Medical Technology leverages advanced IT capabilities to enhance its CRO services.

It is rare in the industry to successfully integrate AI with CRO services. In Hu Qitong’s words, AI is akin to a hammer, while CRO serves as the nail that drives deep into vertical industries. The convergence of this “hammer” and “nail” stems from the perseverance of the Yaoling Medical Technology team. To effectively combine the two, both the hammer and the nail must be robust.

Hu Qitong told VCBeat that he graduated from Johns Hopkins University. As a tech professional with a background in AI, his decision to apply AI to CRO companies stemmed from a chance encounter with Zhang Hongliang, founder of Yaoling Medical Technology.

Before choosing to leverage IT technologies such as artificial intelligence in 2017, Zhang Hongliang had already accumulated extensive experience in the CRO industry and spent many years deeply engaged in clinical trial research. It was precisely because of this background that he identified numerous bottlenecks in clinical trials requiring breakthroughs and transformation, prompting him to establish a company driven by AI to revolutionize clinical trials.

Currently, Yaoling Medical has established collaborations with multiple enterprises and institutions, including Jimin Kexin, Chengdu Better Pharmaceutical, and the Second Xiangya Hospital of Central South University. Moving forward, Yaoling Medical will continue to strengthen its R&D capabilities, aiming to reduce the cost of large-sample real-world studies to less than 1/20 of that associated with traditional models.