Data Intelligence Solution Breaks Through the Bottleneck of Traditional Quality Control Mode of Clinical Trials

In 2016, the British Medical Journal (BMJ) published an article that pointed out that 80% of clinical trials conducted in China had data quality problems. The FDA office in China has said 1,308 of the 1,622 applications should be withdrawn because they contained fabricated, flawed or insufficient data from clinical trials.

But that was 5 years ago, and efforts have been made to address these issues. FDA Office of Scientific Investigations reviewed 2 of the studies from 48 clinical centers. They found that the researchers underreported adverse events and concomitant medications. Follow-up Corrective measures have been taken, including training on good documentation practices." Professor Wu Yilong emphasized that confidence in data quality comes from the clinical trial itself, the so-called "quality by design". (Source: Medicine Cube)

Data quality is not only affected by the clinical trial design itself, but also closely related to the design of the data quality control system. Especially with the acceleration of the globalization of drug research and development, the quality of trial data has become one of the key factors for Chinese research to gain international recognition. Clinical trial data not only affects the results of drug review and approval, but also affects the safety of public medication, and has become a key factor affecting the marketing and application of drugs.

In order to further improve the quality of clinical research in China, in 2015, the former State Food and Drug Administration (CFDA) issued the "Announcement on Carrying out Self-inspection and Verification of Drug Clinical Trial Data", and the former CFDA Audit and Inspection Center (CFDI) launched the Special inspection of research data. After this special inspection, routine inspection of drug marketing registration research based on risk has become the norm.

In June 2017, the former CFDA officially joined the International Council for Harmonization of Technology for Registration of Pharmaceuticals for Human Use (ICH), becoming the 8th regulatory body member in the world. This means that China's drug regulatory authorities, pharmaceutical industry and R&D institutions will gradually transform and implement the highest technical standards and guidelines for international pharmaceutical R&D and production.

In July 2020, the new version of "Good Clinical Practice for Drugs" (GCP) came into effect, requiring investigators to ensure that all clinical trial data obtained from the source files and trial records of clinical trials are accurate, complete and readable. and timely. Source data should be attributable, legible, contemporaneous, original, accurate, complete, consistent, and durable.

In September 2021, the National Health and Medical Commission issued the "Administrative Measures for Medical and Health Institutions to Conduct Researcher-Initiated Clinical Research (Trial)", pointing out that medical and health institutions should establish a management system for clinical research source data, achieve centralized and unified storage, and ensure clinical research. The authenticity, accuracy, integrity, standardization and confidentiality of data in the process of collection, recording, modification, processing and storage, to ensure that the data can be queried and traceable.

The release of this series of policies and regulations has put forward higher requirements for the quality of clinical trial data in my country. On the one hand, the design and execution of clinical trials should be further improved; on the other hand, data quality management should be continuously strengthened. Especially in the context of the explosive growth of the number of clinical trials in China, as a key link in the quality control of trial data, hospital institutions urgently need to optimize the data quality control system and gradually improve the efficiency of quality control.

At this stage, facing the increasing number of trials year by year, due to limited human resources and uneven quality control experience, hospitals and institutions still generally adopt the traditional quality control method—sampling and checking a large amount of data to evaluate the overall test quality. However, this traditional quality control model is highly dependent on manual labor, and it is difficult to fully discover data problems. At the same time, the efficiency of tracing the source is low, and it is generally difficult to objectively quantitatively evaluate the data quality. It has become a bottleneck restricting organizations to improve the efficiency of data quality control.

In order to empower the quality control of clinical trial data, Happy Life Technology (HLT) proposes a digital quality control solution. Under the premise of obtaining authorization, through in-depth analysis of clinical trial source data, combined with lightweight algorithm services, AI technology is used to conduct test data analysis. Comprehensive verification, accurate traceability, and quantitative evaluation are not only convenient and safe to operate, but also significantly improve the efficiency of test data management.

Combined with the quality control experience of leading hospitals, the digital quality control solution proposed by HLT has filled the gap in the industry. It has served many leading hospitals and institutions in just over 2 years since its launch, and carried out digital quality control on more than 40 clinical trials, winning the hospital’s reputation. Unanimously recognized by the institution and the sponsor, and helped the sponsor to successfully meet the national bureau and FDA verification. On the basis of the gradual improvement of hospital informatization construction, more and more hospital institutions and sponsors are also actively taking digital quality control as their tentacles to participate in and promote the gradual progress of clinical trials towards intelligence.

Powered by AI technology, digital quality control can give full play to its unique advantages such as rapid verification, accurate traceability, and quantitative evaluation, further improving the level of data quality management in hospitals. Taking a phase III, randomized, double-blind, placebo-controlled, multi-center study of a tumor clinical trial as an example, from February 2018 to August 2021, the study enrolled 106 patients in both departments A and B of the hospital at the same time. subject. Through the 4-day digital quality control service, a full-scale digital quality control was performed on 2,674 adverse event records and 894 combined medication records. It was found that 872 of the combined medication records were entered correctly, 22 were incorrectly entered, and 183 were underreported; In the adverse reaction records, 2515 were entered correctly, 159 were entered incorrectly, and 40 were missed.

As can be seen from the AECM event underreporting timeline, the problem data found by digital quality control peaked in March 2018. Combined with the analysis of the trial progress, it is found that the peak of the problem data is the peak period of enrollment, and it is also the time node of CRA adjustment. quality.

In addition, through further analysis of the digital quality control results of subjects enrolled in different departments, it was found that the underreported data of department B was obviously more than that of department A. In the case of the same clinical trial design and the same number of enrolled subjects, the quantitative results of digital quality control were verified with the objective situation, and it was found that subjects in department B had more adverse events and more complicated medical records, which led to the occurrence of CRC in the process of data entry, the processing difficulty increases, and false negatives are prone to occur.

With the assistance of digital quality control, hospital institutions can discover problematic data in a very short period of time, at the same time complete traceability quickly with high quality, improve the input data in a timely manner, and optimize the management measures for the trial in the later stage. By accurately and comprehensively identifying the risk of test operation, and evaluating the quality of test execution objectively and quantitatively, the hospital institution has reduced the execution risk of the test from the source, and successfully met the inspection of the National Bureau and the FDA.

With the continuous improvement of the informatization level of medical institutions across the country, multi-center clinical trials have a digital quality control foundation in many centers. The digital quality control service utilizes the standardization, consistency, stability and other characteristics of the algorithm to achieve a unified digital quality control execution standard, quickly find problems and accurately trace the source, and comprehensively and objectively evaluate the key risks in the execution of the test. From one to many, from one trial to multiple trials, from quantitative change to qualitative change, digital quality control drives the entire industry to break through the bottleneck of data quality, and takes a solid step towards digital technology enabling clinical trials.