Testing a new drug is a slow, expensive, and labor intensive process. Artificial intelligence has the potential to reconstruct every stage of the clinical trial process – from matching eligible patients to clinical trials to monitoring compliance and data collection. In the United States, more than 1.7 million people were diagnosed with cancer for the first time in 2018. At the same time, more than 10,000 clinical trials will each require the recruitment of thousands of new patients for testing experimental cancer drugs that may save lives. However, less than 5% of cancer patients eventually participated in the trial. The reasons for clinical trials failing to match the right patient are multiple. For a terminal illness such as cancer, a patient can participate in a drug test only if the existing treatment has failed. Most importantly, not all patients diagnosed with terminal illness are eligible for clinical trials, and even finding them eligible to participate is a daunting task. For those who are eligible to participate in the trial, participating in the trial requires a lot of cost and time, and existing data collection methods make the process more complicated. The process of existing clinical trials is not only to make patients feel sad, but clinical trials are also an inefficient process for other stakeholders such as pharmaceutical companies and CROs. Drug trials took an average of nearly a decade and cost billions of dollars. Many trials failed due to registration issues. These are all indications that this $65 billion clinical trial market needs to be retrofitted. Artificial intelligence is now touted as a panacea, a technology that has great potential to simplify the cumbersome clinical trial process. For example, from remotely monitored IoT to machine learning for EHR data, to artificial intelligence-based data protection for network security. CBinsight, a well-known data service agency in the United States, recently analyzed the prospects and limitations of artificial intelligence in clinical trials. The arterial network compiled and compiled this.  The main contents of this paper include: 1. Why do pharmaceutical companies need to speed up the drug development process? Current status of clinical trials 2. How does AI change the clinical trials? Patient recruitment on matching patients and trials Streamline the registration process Data collection and dependency 3. Apple is preparing to reconstruct clinical trials 4. Why only AI can't be a panacea Why pharmaceutical companies need to speed up the drug development process The launch of new drugs is a long and arduous process. Although there is no central repository of expenditures and timelines for drug development, the study estimates that it takes an average of 7.5 years for new drugs to be tested on patients, with costs ranging from $161 million to $2 billion per drug. It is well known that clinical trials need to be carried out in multiple stages, and in this process, costs and costs are incremental. Phase III trials require more patient populations and are much more expensive and complex than Phase I trials. Even if a lot of time and money have been invested, in the end, only one-tenth of all new drugs that have undergone a phase of clinical trials will be approved by the FDA. There are many reasons for failure in clinical trials, including failure to recruit enough participants, midway patient withdrawal, accidents and serious side effects, and improper data collection methods can lead to the death of new drugs. Of course, the more clinical trials that fail in the later stages, the more the clinical trial initiators and patients pay a higher price. Sometimes the blow to clinical trial failure is likely to be fatal for pharmaceutical companies. For example, Novartis Switzerland attributed its 15% decline in net income in the first quarter of 17 years to a failure of a phase III drug for the treatment of heart failure. This is nothing, in the United States, the pharmaceutical company Tenax Therapeutics' main drug failed two months after the third phase of the trial, the CEO resigned. According to reports, the company is considering a merger or sale. For small biopharmaceutical and start-up pharmaceutical companies entering the field, the cost of failure is even more significant. The risk of failure is absolute, especially if the pharmaceutical company only bets on a drug. The high costs associated with drug development will ultimately be passed on to consumers. Consumers' purchases of health care services not only include the cost of successful drug development, they also need to pay for failed clinical trials. This is why the price of medicine is so high. Except for economic costs, the cost of failure or ineffective trials is the patient. For most patients, finding clinical trials is an attempted process, but registration and participation pose new challenges. Many clinical studies still use raw and outdated methods to collect and validate data, send patient medical records by fax, manually count the remaining tablets in the bottle, and rely on patient diary records to determine medication adherence. Obviously this process needs to be refactored. How AI will change clinical trials It can be said that AI has the ability to change every aspect of clinical trials, from clinical trial registration to patient compliance management. Discovery clinical trial By changing the existing clinical trial process, matching the right patient to the required clinical trial can save a lot of time, both for the patient and for the pharmaceutical company. According to Cognizant's report on recruitment forecasts, approximately 80% of clinical trials failed to recruit appropriate patients within the timeframe, and approximately one-third of Phase III clinical trials were terminated due to patient recruitment difficulties. Currently, more than 18,000 clinical studies in the United States are recruiting patients, of which only about 1,000 are breast cancer. CBinsight quoted the White House's statement in May 2018: "In fact, only 3% of cancer patients can participate in clinical trials." If the doctor knows about the ongoing trial, the patient may occasionally get a trial recommendation from the doctor. Otherwise, patients will need to find relevant information from the government department ClinicalTrials.Gov. ClinicalTrials.Gov is a comprehensive federal database that collects past and ongoing clinical trial information. If there is AI, the ideal solution is to use artificial intelligence software to extract relevant information from the patient's medical record, compare it with ongoing trials, and recommend matching studies. In fact, extracting information from medical records—including EHRs and laboratory images—is one of the most popular applications of artificial intelligence in the medical field. However, the real landing of this technology has to face some challenges, such as unstructured medical data and different data sources that do not communicate with each other.  EHR Interoperability Challenge Although the federal government has spent more than $28 billion over the past decade to implement digital electronic health records (EHRs), there is no centralized repository or standard format for patient medical data. In fact, it is difficult for patients to integrate all medical information from all the medical institutions he has visited. The HIPAA Act is a law that protects patient medical data and privacy. It allows patient data to be shared with personally identifiable information such as name and SSN number, based on the patient's signature of informed consent. This allows AI startups to analyze these medical data in minutes and recommend eligible patients to follow the current process, which will take several months. However, the problem of interoperability (compatibility) still exists, and it is not easy to share health information between organizations or different software. Different hospitals treating the same patient may use different EHR software to enter data. In clinical trials, researchers will still send fax requests for specific patient records to different hospitals, which will then fax or send the information in the form of pictures (including handwritten notes or pictures of PDF files). This poses a challenge to artificial intelligence technology. A study by researchers at the Massachusetts Institute of Technology, Harvard University, Johns Hopkins University, and New York University showed that "in clinical annotation, standard natural language processing tasks such as emotional analysis and word disambiguation are difficult. Because clinical data often has spelling errors, acronyms, and lots of copy and paste." Health artificial intelligence company Flatiron Health further explains this in a patent application: "Structured data may also become unstructured due to transmission methods. For example, spreadsheet faxing or conversion to read-only documents (such as PDF) will Lose most of its structure. This outdated artificial system makes it difficult for clinical trial researchers to collect the accurate data needed to determine patient eligibility. Last year, the California-based startup Mendel's AI tried to solve this challenge by integrating the medical history of cancer patients, who can submit their medical records on Mendel's platform. Alternatively, the patient can allow Mendel to collect all medical records from the doctor on their behalf. Mendel is developing machine learning algorithms that extract information from digital records and match patients to the tests that are best suited to their needs. The startup charges patients for subscriptions and processes all medical records at the starting price of $99 for the first three months. However, Mendel has not announced further expansion plans since then and has not raised any further funds. In addition, the option to upload medical records on their website is no longer available. Other companies, such as Antidote.me, are using machine learning to simplify terminology in the "inclusion/exclusion" standard listed on the ClinicalTrials.Gov website. In B2B, startups are using deep learning and natural language processing to automate clinical trial matching by working directly with health agencies. For example, Deep 6 AI works with Cedars-Sinai Medical Center and TD2. TD2 is a cancer CRO company.  Acquisition becomes a strategy to solve interoperability In 2014, Flatiron Health acquired Altos Solutions, a cancer-focused EMR company, to address interoperability issues. At the time, Flatiron was selling its cloud analytics platform to medical and life science companies. Altos' EMR is also used by the Florida Cancer Specialists oncology facility. The deal gives Flatiron direct access to the patient's raw data, rather than relying solely on third-party email. To date, more than 2,500 clinicians have used Flatiron's OncoEMR, and related reports indicate that more than 2 million active patient records are available for research. In addition, Roche acquired Flatiron for $1.9 billion in February 2018. This has become one of the largest M&A deals in the field of artificial intelligence. This, coupled with Roche's recent acquisition of other digital initiatives such as Foundation Medicine and Viewics, underscores Roche's intention to reposition itself as a machine-driven, high-profile pharmaceutical company. Clinical trial registration challenge The registration process does not end when the patient finds a suitable clinical trial. Instead, the patient must visit the place where the trial is conducted to confirm his eligibility and the patient is required to conduct a preliminary telephone screening with the clinical research team's researchers. Patients must also meet the inclusion and exclusion criteria in the trial list, and each patient must qualify for a study. These standards are often filled with medical terms. For example, in a clinical trial of breast cancer approved by the FDA, patients must pass assessments such as laboratory and imaging tests to ensure she meets all inclusion and exclusion criteria. Depending on their availability and their distance from the test site, some patients can complete these procedures in less than a week. But for those who are busy with work, parents or long commutes, this process may require multiple business trips. As part of the qualification, the site investigator collects the patient's medical records from other doctor's offices. As mentioned earlier, these fax or email copies add a layer of complexity when extracting information using artificial intelligence. If the conditions are met, the patient signs a consent to agree to the terms of the clinical trial. This includes recognizing potential side effects, willing to provide biological samples containing genetic information, and paying for expenses not included in the research budget. Some solutions that use artificial intelligence to extract information from patient records can help simplify the registration process by automatically verifying certain inclusion and exclusion criteria. A more effective solution is to use patient-generated data. Due to the large amount of data generated in real time, AI and patient-generated data are inseparable. Apple is exploring applications in this area. Apple is building a clinical research ecosystem around the iPhone and Apple Watch. By continuously monitoring patients in a comfortable home, the company can generate a large amount of health data that was previously unavailable. Apple has been working with medical institutions and medical researchers to easily identify the patient population that is right for their research. This section will be explored in detail below. Clinical trial patient compliance After the patient confirms to participate in the clinical trial. When the patient returns home, he will take the first course of medication (for example, a 30-day vial to indicate the dose) and the patient diary that needs to be filled out each day. Many clinical studies still use paper diaries rather than electronic systems. When taking a study drug, patients need to pay attention to other drugs taken at the time, as well as any adverse reactions (including headache, stomach pain, muscle pain). It can be seen that the recording of the entire clinical trial data relies on unproven sources of information, such as patient memory and paper records, which can lead to inefficiency. Dependence on patient memory: When patients are on regular checkups, the researchers examine their vials to ensure that there are no remaining pills and to check for gaps or inconsistencies in the patient's diary. If there is a lack of information in the diary entry, the researcher relies on the patient's memory of the event. This makes this process prone to human error. Outdated recording system: Paper documents may lose information, which is an outdated and unreliable way, and the key information it records may be lost. Patients give up risk: frequent visits to clinical research sites for regular check-ups can put time and money on the patient, especially for those who travel abroad (tickets, hotel stays, vacation time, etc.). This increases the risk of patient suspension. Additional fees: Although the patient signed a consent form that included out-of-pocket expenses, many patients did not understand that there might be additional costs. For example, additional MRI and laboratory tests performed during follow-up may not be included in the trial, and insurance companies will not cover these costs because they are used for "research purposes" rather than "medical needs." How AI can help with medication and improve compliance Failure to adhere to treatment, or not taking the drug in the right way, may adversely affect the health of the patient. For a researcher, if a study recruits a new patient, it will increase costs and affect the accuracy of the results. The initiators of clinical research are eager to change this situation and invest in technology for this. In addition to using mobile technology to remind patients to take medicine, Pfizer and Novartis also invest in the Internet of Things and “absorbable sensors†to track drug intake. In the first quarter of the year, Merck Ventures participated in Medisafe's $14.5 million Series B financing, which developed wireless vials. Artificial intelligence start-ups go a long way in patient compliance and provide a visual confirmation program. AiCure, a New York-based mobile SaaS platform, uses image and facial recognition algorithms to track patient compliance. The patient recorded a video of his own swallowed tablet on the mobile phone, and AiCure confirmed that the correct person had taken the correct pill. AiCure has raised a new round of financing of $27 million after a series of funding from institutions such as the National Institutes of Health and the National Institute on Drug Abuse. Another startup in this space, Catalia Health, backed by Khosla Venture, is developing a medical partner and consultant using artificial intelligence. Catalia hopes to enhance patient behavioral changes by asking specific questions, setting reminders, and tailoring conversations for each patient. One of the company's goals is to understand why patient compliance is poor. Catalia's robotic assistant is essentially a tablet with a touch screen, but it is reported that the startup is also working on voice activation. It is worth noting that the ability of artificial intelligence assistants to successfully change lifestyles depends to a large extent on whether patients are willing to interact with artificial intelligence every day – this can be monitored through artificial intelligence IoT technology.  AI+ IoT provides continuous daily feedback An emerging trend is to integrate biosensors with artificial intelligence. To this end, some startups are developing their own monitoring devices and sensors, and then adding a layer of machine learning to interpret the data. Others are only developing artificial intelligence software and integrating with third-party home monitoring equipment. Israeli company ContinUse Biometrics is developing its own sensing technology. The startup monitored more than 20 biological parameters—including heart rate, blood glucose levels, and blood pressure—some of the standard life characteristics measured during drug testing—and used artificial intelligence to detect abnormal behavior. The company raised $20 million in Series B financing in the first quarter of 2018. Biofourmis is developing an artificial intelligence analysis engine that extracts data from FDA-approved home monitoring devices and predicts patient health outcomes. AI + Internet of Things has the potential to monitor patients' physiological and behavioral changes in real time and in clinical trials, potentially reducing the cost, frequency and difficulty of on-site inspections. How Apple is preparing to reconstruct clinical trials For small start-ups, the biggest obstacle to simplifying clinical trials is that these technologies are relatively new and the industry adapts slowly. However, technology giants such as Apple have seen success in attracting partners to participate in their health care initiatives. Apple has been addressing some of the bottlenecks in the medical industry's information flow, allowing researchers to use their APIs to build artificial intelligence applications. In particular, Apple is building a clinical research ecosystem around its two devices, the iPhone and Apple Watch. Data is at the heart of artificial intelligence applications. Through these products, Apple can provide medical researchers with two data streams, both of which are currently difficult to obtain. Apple's big data stream Since 2015, Apple has launched two open source frameworks, ResearchKit and CareKit, to help clinical trials recruit patients and remotely monitor their health. These frameworks allow researchers and developers to create medical applications to monitor patients' daily lives. For example, researchers at Duke University have developed an app called Autism & Beyond that uses the iPhone's front camera and facial recognition algorithms to automate a child. Screening for symptoms. Similarly, nearly 10,000 people use the mPower app, which uses exercises such as finger tapping and gait analysis to study Parkinson's disease with the user's informed consent. Other variants of the app, such as HopkinsPD for Android users, use machine learning to process data collected from smartphones to assess the severity of the patient's Parkinson's condition.  Refactoring EHR data sharing In January of this year, Apple announced that iPhone users can now access all their medical records through the iPhone's health app. Called the "health record" API, this feature is an extension of the work done by AI-based medical startup Gliimpse before it was acquired by Apple in 2016. Information on the Gliimpse website shows that its "core technology is to turn medical documents into data and easily search and filter the most important graphics and dashboards." This is consistent with Apple's health record function – an easy-to-use interface where users can find allergies, diseases, immunizations, laboratory results, medications, procedures and vital signs they need. Apple also works with popular EHR vendors such as Cerner and Epic to address interoperability issues. One potential end goal of these partnerships may be a two-way data flow, in which case EHR vendors are encouraged to integrate patient-generated data into Apple software. A statement from Apple said: "More than 500 doctors and medical researchers use Apple's ResearchKit and CareKit software tools for clinical research involving 3 million participants, ranging from autism and Parkinson's disease to post-surgery Family rehabilitation and physical therapy." In June of this year, Apple launched a health record API for developers. Users can choose to share their data with third-party applications or medical researchers. This opens up new possibilities for disease management and lifestyle monitoring. What does this data mean for clinical research? Apple is now at the heart of a new medical data ecosystem that provides previously unavailable health data while collecting difficult-to-integrate EHR information. The application scenario will be endless in the use of artificial intelligence and machine learning for early diagnosis, driving drug design decisions, recruiting appropriate patients for research, and remotely monitoring patient progress throughout the study. One of Apple’s potential competitors in this space is Google. Google's Project Baseline program is designed to recruit 10,000 patients and monitor their daily lives for five years, which may ultimately benefit clinical trials. Many trials included an experimental group (patients taking the study drug) and a control group (patients taking a placebo). The purpose of the control group was to establish a baseline to compare with the symptoms of the experimental group. The patient usually does not know what medication he is taking. Patient-generated data—as Google's Project Baseline is collecting—can eliminate the need for control groups, provide the data needed for control, and ultimately reduce patient recruitment difficulties. However, the project is still in its early stages and it may take several years for specific applications to be realized.  Why is AI not only omnipotent? The healthcare industry is a leader in artificial intelligence applications, testing applications ranging from machine-assisted diagnostics to extracting information from electronic health records. In particular, the use of AI to discover new drugs is also growing, with pharmaceutical giant Merck working with startup Atomwise, GlaxoSmithKline and Insilico Medicine to apply AI to new drug development on. However, the application of artificial intelligence in the actual clinical trial process is still in its infancy. There are fewer start-ups directly targeting customers in the clinical trials than in other areas of healthcare. In many aspects of clinical trials, the landing of AI requires, first and foremost, the digitization of the entire industry. As mentioned above, many trials still rely on paper diaries to obtain data. Even if some parts are stored in digital format, it is difficult to search. Handwritten clinical records must overcome the challenge of dealing with natural language processing in extracting information. One of the biggest obstacles in clinical trials will be to overcome inertia and completely update processes that are no longer effective. Another challenge will be to accurately understand where artificial intelligence can help and its current limitations. Researchers must engage in discussions around achievable short-term goals, rather than fantasizing AI to make all problems disappear. For now, developing new data sources and handing them over to patients – as Apple is doing – has proven to be revolutionary in clinical research, and a great example is the use of APP for patients. The Parkinson severity score. Ultimately, the goal of applying artificial intelligence to clinical trials will be to narrow the gap between the information currently available to patients and the information they really need to lead them to a healthier life. Grape seed extract;Astaxanthin;Anthocyanins(Anthocyanins are a powerful anti oxidant, which protects the body from damage caused by a harmful substance called free radicals. Anthocyanins also enhance the elasticity of blood vessels, improve the circulatory system and smoothness of the skin, inhibit inflammation and allergies, and improve the flexibility of joints. );Vitamin C powder Anthocyanin;Grape seed extract;Astaxanthin;Polyphenol Shaanxi Zhongyi Kangjian Biotechnology Co.,Ltd , https://www.zhongyibiotech.com
AI is refactoring the drug clinical trial process, and Apple is working on data reconstruction clinical trials