Advancing Real-World Evidence in China: An Integrated Platform Bridging Hospitals, Pharma, Insurers, and Big Data

Real-world study (hereinafter referred to as RWS), which originated from pragmatic clinical trials, falls within the scope of pharmacoepidemiology.

RWS refers to the evaluation of diagnostic and therapeutic efficacy, safety, and pharmacoeconomics in real-world settings, based on a larger sample size (covering a broader clinical patient population) and treatment measures tailored to patients' actual conditions, treatment preferences, and economic circumstances.

As health and medical big data has become one of the national development strategies, RWS exploration based on big data has gradually become a hot topic in healthcare industry research.

To gain a clear and in-depth understanding of the current development of Real-World Studies (RWS), VCBeat conducted exclusive interviews with numerous industry experts from hospitals, pharmaceutical companies, big data firms, genetic testing companies, and commercial insurance providers. By examining RWS from multiple dimensions, this article presents its most authentic aspects to readers.

RWS is a shared demand among three parties: hospitals and physicians, pharmaceutical companies, and insurance companies.

For physicians, the aspiration is to continuously distill diagnostic and therapeutic patterns from clinical practice, refine clinical management strategies, and provide patients with optimal clinical solutions. This represents the lifelong pursuit of high-caliber clinicians and serves as the sole means by which they can elevate their professional standing, academic stature, and status.

If the market launch of a pharmaceutical product is likened to the theatrical release of a film, then extensive distribution efforts are required after its debut. Furthermore, substantial secondary exploitation of the film’s intellectual property (IP) is necessary to continuously amplify the value of this technological and creative work.

Therefore, for pharmaceutical companies, the launch of a product is by no means the endpoint, but rather the starting point for extensive Real-World Studies (RWS) and clinical application evaluations. After a drug is launched, pharmaceutical companies need to observe its effectiveness, safety, and pharmacoeconomics in populations beyond those included in rigorous Phase III registration trials, thereby extending the drug’s lifecycle, enhancing its commercial value, and raising its market potential ceiling.

It is precisely due to the foundational and platform-level value of Real-World Studies (RWS) for major payers in the healthcare industry that numerous medical big data companies have elevated RWS to a strategic priority and industry-wide prominence.

Liu Lunxu, Vice President of West China Hospital, Sichuan University. According to his recollection, he first heard the term “real-world study” around 2012. At that time, a medical academic conference featured a discussion topic on real-world studies.

In Liu Lunxu’s view, real-world studies (RWS) require the standardization of existing clinical workflow data. RWS is not a specific project but rather data generated through long-term accumulation of real-world practice. To effectively leverage these data, standardization must be implemented throughout the processes of data generation and accumulation.

Real-world studies (RWS) can reflect the most objective reality of patient treatment conditions, diagnostic and therapeutic strategies, and outcomes. They differ significantly from rigorous randomized controlled trials (RCTs). RCTs represent highly specialized scenarios, whereas the complexity encountered in RWS far exceeds the diversity seen in RCTs. Therefore, in terms of clinical guidance, the results of RWS are more valuable.

Three key elements are critical to Real-World Studies (RWS): First, big data. RWS must be grounded in big data to yield effective guidance. Second, data standardization. The generation of big data requires standardization, including the use of standardized clinical terminology; otherwise, the accuracy of data analysis cannot be ensured. Third, unified collaboration. While RWS application scenarios may involve a single department or hospital, implementing them on a larger regional scale or nationwide necessitates concerted efforts from all stakeholders.

ZeroKrypton Technology’s research analysis platform is a big data platform deployed at West China Hospital, designed for real-world studies (RWS) in oncology.

From the current perspective, there are two major drivers for hospitals to engage in Real-World Studies (RWS). At the research level, hospital promotion systems incentivize physicians to publish papers. However, as the volume of published papers increases, it has become increasingly difficult to publish articles with high academic impact.

RWS can assist physicians in conducting more in-depth scientific research, thereby facilitating the generation of research outputs and the publication of high-impact papers. Experts aim to cultivate their influence within a specific field through focused research endeavors.

Additionally, the state has established special funds for Real-World Studies (RWS). By undertaking these studies, hospitals not only secure dedicated funding but also leverage their research findings to inform and facilitate the formulation of healthcare policies by health administrative authorities.

“Not all physicians participate in real-world study (RWS) research; such studies are more commonly conducted by medical teams within a department. A medical team comprises the department head, attending physicians, and other team members. The department head grants tiered data access to subordinate physicians through a big data platform. Generally, resident physicians and interns do not participate,” said Liu Lunxu.

For pharmaceutical companies, Real-World Studies (RWS) represent a mandatory requirement. For instance, national regulations explicitly mandate that pharmaceutical companies must submit drug safety monitoring data within five years after a drug’s market approval; failure to do so risks the drug’s withdrawal from the market. RWS serves as an effective approach for pharmaceutical companies to meet such regulatory obligations.

The ultimate application of Real-World Studies (RWS) is to determine, through insights derived from real-world data, whether a drug can be extended to broader patient populations to enhance patient benefits. An industry practitioner told VCBeat, “Traditionally, pharmaceutical companies have had to rely on costly Randomized Controlled Trials (RCTs) and endure lengthy timelines to identify new indications. This process not only involves substantial registration costs but also carries significant risks. In contrast, RWS enables pharmaceutical companies and experts to conduct preliminary analyses and initiate trials earlier.”

Generally, physicians determine the scope of drug use based on the product labeling. In practice, off-label use is common in many hospitals, where physicians prescribe medications based on their clinical judgment and experience. This phenomenon is not uncommon across the pharmaceutical industry. By leveraging Real-World Studies (RWS) and conducting effectiveness analyses, pharmaceutical companies can gather evidence to decide whether to expand the indicated indications for their products, thereby broadening the therapeutic applications of their drugs.

On November 12, Novartis Pharmaceuticals released real-world study data from Germany involving patients with heart failure with reduced ejection fraction (HFrEF) who were treated with Entresto® (sacubitril/valsartan). The data further support the beneficial effects of Entresto on heart failure symptoms and quality of life in patients, consistent with findings from the PARADIGM-HF trial and other real-world study cohorts. These data were presented at the 2017 American Heart Association Scientific Sessions.

Shreeram Aradhye, Chief Medical Officer and Global Head of Medical Affairs at Novartis, stated, “Patients with heart failure experience symptoms that severely impair physical activity and quality of life. A growing body of evidence demonstrates that Entresto improves the quality of life in patients with heart failure while reducing cardiovascular mortality and hospitalization rates for heart failure.”

This non-interventional, retrospective dataset study examined changes in clinical characteristics over one year following the initial dosing regimen among 1,643 patients in Germany who were prescribed Entresto. All available data were analyzed to derive the corresponding study findings. (For detailed results, please refer to the Sina Medicine translation:Real-World Data: Novartis’ Entresto Improves Quality of Life in Heart Failure Patients）。

Zhao Cen, currently Deputy Secretary of the Party Committee and Vice Dean of PBC School of Finance, Tsinghua University; Director of the China Insurance and Pension Research Center; Deputy Director of the Internet Finance Laboratory; and Standing Member of the National Financial Youth Federation.

According to Zhao Cen, the Insurance Center of Tsinghua University PBC School of Finance was established in November 2016, and its research on big data and product innovation is based on RWS.

The development of insurance products often requires real-time, standardized, and secure data sharing. Throughout the entire commercial loop of health insurance, data is needed not only during the product design and development phase but also to support later-stage health interventions and health management processes.

“We hope to research and develop innovative insurance products while applying advanced big data technologies to product management, which makes real-world data more meaningful. For instance, in product pricing, precise pricing can only be achieved by using actual data on medical costs and incidence rates,” said Zhao Cen.

Specifically, insurance companies leverage deep data mining to first break away from traditional actuarial models in pricing, enabling precise premium determination based on individual medical conditions, credit histories, and other factors. Secondly, during the underwriting and claims settlement stages, big data technologies are employed for fraud detection. Furthermore, after policy issuance, insurers predict disease progression based on health indicators and lifestyle factors, shifting from post-claim compensation to pre-emptive intervention. This approach facilitates proactive health management, achieving a win-win outcome.

According to Zhao Cen, future data sources for the Insurance Center may include traditional institutions such as healthcare providers, insurance companies, and specialized medical big data firms; internet-based sources such as social media data; smart hardware and wearable devices; and government agencies such as the Ministry of Human Resources and Social Security. LinkDoc Technology, as a key partner, is also deeply involved in this initiative.

As the founder of Zero2IPO Tech, Zhang Tianze first connected with RWS around early 2014. At that time, there were already emerging voices and perspectives internationally suggesting that randomized controlled trials (RCTs) imposed overly stringent controls on the experimental process, thereby limiting the strength of their evidence. Due to limited sample sizes and idealized control of variables, RCTs often exhibited varying degrees of discrepancy between real-world drug utilization patterns and clinical study findings when applied to larger samples, more diverse populations, and more complex clinical scenarios.

This event opened a new perspective for Zhang Tianze in viewing clinical data.

As one of the earliest partner hospitals of LinkDoc Technology in data processing, Professor Zhang Xun from Tianjin Chest Hospital also accepted an interview with VCBeat. Notably, Professor Zhang was listed among the “2016 Highly Cited Chinese Researchers”—a roster comprising scholars with world-class influence in their respective fields of research.

In his daily practice, Professor Zhang employs extensive real-world data to evaluate treatment regimens. He has noted in several papers that discrepancies exist between patients’ real-world clinical treatment data and findings from pharmaceutical research, and these discrepancies highlight the potential for further value extraction from drug development.

“As reflected in Professor Zhang’s research findings, drugs such as aspirin have a long lifecycle, yet we continue to uncover their new medical value, all of which requires discovery through real-world data,” said Zhang Tianze.

Zhang Tianze believes that physicians conducting real-world studies (RWS) need to overcome two major challenges: first, organizing and documenting patients’ clinical medical records into phenotypic data suitable for statistical analysis; and second, tracking patients to determine their survival duration, survival status, and quality of recovery.

In many hospitals, data collection and organization are carried out by graduate students, doctoral candidates, or junior physicians. Due to their limited clinical experience, the quality of the compiled data is often inconsistent, with numerous errors and omissions.

Patient follow-up is often initiated only when physicians have research needs, with nurses or doctors from the department making individual phone calls. By that time, however, very few patients can be reached. Consequently, the statistical evaluation metrics generated in these two scenarios fail to meet the requirements of clinical research. “LinkDoc’s starting point is to help hospitals conduct real-world studies at a research-grade level,” Zhang Tianze told VCBeat.

As a conceptual research approach, the practical implementation of RWS in hospitals, pharmaceutical companies, and insurance firms is no easy feat.

According to a clinical medicine expert at LinkDoc Technology, real-world studies (RWS) are characterized not merely by large data volumes but, more critically, by their longitudinal nature. Therefore, LinkDoc Technology provides an end-to-end integrated solution covering all aspects of the RWS process. This closed-loop product encompasses hospital data collection, data structuring, post-discharge patient follow-up, and doctor-patient communication, with modules designed for physicians, patients, and follow-up personnel. These systems share highly interconnected underlying data on the LinkDoc platform, linking physicians, patients, and research stakeholders to ensure the reliability of RWS data.

The main challenges in implementing this system fall into three categories:

First, how to provide high-quality data;

Second, new statistical methods;

Third, long-term, continuous, and in-depth patient follow-up.

In previous study design research, physicians already had clear ideas for experimental protocols, akin to product mold-making, where doctors used predefined models to collect data. LinkDoc’s real-world study (RWS) approach is more comparable to 3D printing, proactively preparing diverse data sets for clinicians so they can evaluate drugs and treatment regimens comprehensively from any perspective with ease. Achieving high data quality alongside extensive richness and completeness simultaneously is challenging; this is precisely why LinkDoc spent three years building a team of over 700 professionals to tackle data-related challenges in the field of oncology.

Furthermore, LinkDoc must design distinct research topics tailored to clients’ products and business needs. Whether the focus is on therapeutic efficacy data or treatment safety data, LinkDoc’s team is required to perform high-quality data structuring for various elements within clients’ research protocols, including treatment durations associated with different drugs and therapeutic regimens.

“After all, our clients ultimately need to be able to translate their findings into publishable articles or establish new standards for diagnosis and treatment,” the expert told VCBeat.

Furthermore, the volume of data in Real-World Studies (RWS) must be sufficiently large. In contrast, representative Randomized Controlled Trials (RCTs) typically involve small sample sizes; for instance, in oncology research, a few hundred cases are generally considered a large study sample. According to expert estimates, data from at least 100 Grade III Class A hospitals are required to efficiently support the conduct of RWS. Traditional Electronic Data Capture (EDC) systems struggle to handle such massive datasets, necessitating specialized software and technological solutions for adequate support.

Currently, LinkDoc has covered hundreds of hospitals in China.

In terms of real-time data, hospitals need to transform clinical records from unstructured electronic medical records or even paper-based charts into structured formats. Regarding long-term outcomes, hospitals must implement long-term follow-up, which is a systematic project involving substantial workload.

Randomized Controlled Trials (RCTs) feature rigorous methodological designs and strict quality control throughout the process. In contrast, Real-World Studies (RWS) inherently possess the advantage of abundant data resources and massive data volumes. Consequently, the design of research protocols in RWS involves a wider array of variables and more diverse methodological options. This places higher demands on data processing and statistical analysis. Therefore, RWS presents new challenges and requirements for data processing and statistical methodology.

Since patient data encompasses both in-hospital and out-of-hospital sources, patient management requires a dedicated team to systematically manage medical records collected within specific timeframes. Physicians need to continuously monitor changes in patients’ conditions and observe survival outcomes through regular follow-ups.

Solution:

In fact, what the real-world study (RWS) industry currently lacks is not merely a robust system. Since RWS research requires extensive manual involvement, having a system alone is insufficient if physicians lack the time to implement it, thereby failing to achieve the desired outcomes. Therefore, LinkDoc has built upon this foundation by engaging dedicated professionals throughout the entire process, from data collection to subsequent patient management.

To address technical challenges, LinkDoc has deployed three types of statistical analysts. One type comprises analysts with a research background, particularly in randomized controlled trials (RCTs). These analysts possess a deep understanding of scientific research and, by integrating physicians’ insights, can effectively help them extract target data.

The second category comprises algorithmic statisticians within technical teams. Their strength lies not in their understanding of medical data, but in providing researchers with diverse recommendations at the algorithmic level. Although algorithmic statisticians lack a medical background, they excel at identifying correlations within data and can offer researchers valuable insights.

The third category is application-oriented data analysts. These analysts previously worked in strategic planning for companies. Leveraging their understanding of data and industry insights, they can translate data into practical applications. In simple terms, application-oriented data analysts transform real-world data into actionable products, thereby guiding the operations of insurance institutions or corporate marketing departments.

Furthermore, for clinical data to be suitable for scientific research, it must meet high-quality standards; therefore, structuring the data through natural language processing and machine learning is essential. Additionally, a data entry team is required for data annotation. To ensure that data entry rules align with hospital documentation practices, quality control and medical teams are necessary to guarantee that all data comply with medical requirements.

Furthermore, given the substantial volume of follow-up required after patient discharge, it is essential to establish a sufficiently large and professional follow-up team.

Currently, LinkDoc’s medical services team has grown to over 400 members, including a follow-up team of more than 200 professionals. Service personnel at traditional health IT companies rarely possess the aforementioned capabilities, which constitutes the most significant distinction between LinkDoc and traditional health IT vendors and channel partners.

“It is not easy to achieve a reasonable return on investment (ROI) for real-world studies (RWS) at both the individual project and industry-wide levels,” stated Zhang Tianze. The costs of RWS primarily consist of four aspects:

1. Prepare a vast repository of social data resources;

2. Establish, cultivate, and develop a professional team specializing in medical affairs, data management, and statistical analysis for cross-disciplinary Real-World Studies (RWS);

3. Developing the corresponding technical software platform requires extensive data for iterative processing and refinement;

4. Academic services for hospitals and physicians should prioritize empowering them to effectively utilize data, as this is the starting point for enabling data flow and mining.

According to Zhang Tianze’s understanding, investing in RWS is akin to building a road: once the road is constructed, vehicles transporting fruits and vegetables, coal, and oil can all use it. If the road is built solely for transporting vegetables, its costs would certainly not be recouped.

However, if this pathway can be reused by the entire industry, its costs will not only be recouped but also yield substantial returns.

This is precisely why real-world studies (RWS) must encompass multi-domain, interdisciplinary, and cross-industrial reusability. Without integrated design from the outset, RWS cannot achieve the goal of “clean once, use a hundred times.” Undoubtedly, under such circumstances, cost recovery would be difficult.

The infrastructure serves as the foundational platform for Real-World Studies (RWS), while its supporting facilities and service capabilities encompass the software, technology, and medical data management, statistical analysis, and interpretation inherent to RWS. Constructing this infrastructure is a multidisciplinary collaborative effort that requires proactive, forward-looking planning.

It is difficult for a single company or a few companies alone to bear the cost of Real-World Studies (RWS). However, if 30% of the industry’s enterprises collaborate to share the costs and utilize RWS data to evaluate drug therapeutic efficacy, the value of the data can be amplified tenfold or even a hundredfold. From this perspective, despite the upfront investment, the returns from RWS will be substantial.

In response to this claim, VCBeat also consulted Jin Ge, CEO of Yuheng Gene, a subsidiary of Yuheng Pharmaceutical. According to her, the enrollment cost for a single clinical patient is approximately RMB 300,000.

Under normal circumstances, the approval of a specific category of drugs requires the enrollment of thousands of participants. The development of a biologic drug, spanning from research and development through Phase I, II, and III clinical trials, often necessitates billions of dollars in funding before it can receive final approval from the National Medical Products Administration (NMPA). If big data companies such as LinkDoc Technology could leverage Real-World Studies (RWS) to accelerate this process, reduce costs for pharmaceutical companies, and improve drug conversion rates, these companies would be highly willing to pay for such services.