Frequently asked questions
What is the big problem with clinical trials?
While clinical trials are the only way to access new drugs and therapies before final FDA approval, patients who lack alternatives often have immense difficulty meeting inclusion and exclusion criteria, the complex protocol requirements which must be met to enroll and participate in a trial. Patients often are facing poor results of first lines of therapy as they seek new alternatives. ~98% of physicians are not running clinical research trials and are not aware of the latest therapies under study. As a result, patients have limited access, time, and information on how they can participate.
Academic institutions have teams of clinical research administrators manually reading electronic medical records to scan for patient histories and current statuses that would qualify them for the trial. This manual process to find patients that fit the criteria can take up >6 months, costing significant time and money, while limiting the window of accessibility to patients who might fit the criteria. As a result, roughly 2/3 of trial enrollment volumes are left unfilled. Historically, trial protocols have continued to require manual subjective review to identify patients. Dyania Health has launched a software that uses natural language processing to understand and extract streamlined criteria from written text in clinical notes (biopsy, medical histories, radiology).
This allows, for the first time, the automation of identification of patients with a much higher accuracy on subjective criteria than just on laboratory data alone. It also allows for an instantly larger reach to volumes of patient populations who might not have a physician aware of latest trials. Dyania Health’s software acts as an anonymized search engine for finding patients across health care systems. It can be integrated on a per trial site basis, anonymizes and redacts all HIPAA patient identifiers, and extracts the clinically relevant information needed for research and identifying patients for trials.
Why is Dyania Health the right company to solve this problem?
Dyania Health is comprised of a team of clinicians and software engineers who are pharma veterans, serial drug inventors, and principal investigators and have been at many of the top global institutions including Harvard, UCLA, Bayer Pharmaceutical, and IQVIA. The executive team combined has more than 600 academic publications and 120 years of experience working on clinical trials, drug development, and clinical research. This company expertise allows for an immensely higher accuracy, a deep respect for compliance, and a reliable enterprise solution.
How does compliance work?
Dyania Health maintains a patient-first approach to data privacy with the highest respect for HIPAA and GDPR compliance and leaving hospitals with complete control over their own confidential patient identifying information. In addition to excluding all patient identifiers during the anonymization process Dyania Health’s software includes a scrubbing algorithm that automates the redaction of identifiers within clinical text. The algorithm even utilizes syntax to understand that “Mr. John Tumor” is referring to a patient name and not a clinically relevant characteristic of oncology.
How are patients matched if they are anonymized in the database?
Dyania Health maintains as a core company philosophy that patients should always hear about and discuss any opportunities for a trial with their own physician in person. A patient’s own physician knows them best and will have the best knowledge of a patient’s intricate medical history, expertise on whether the specific patient could benefit, and strong judgement on whether the patient would be likely to adhere to the trial.
Each trial site healthcare system maintains control over their own anonymization secret database which replaces Patient ID numbers with randomized Research ID numbers. This mapping is used by the specific healthcare system to find and re-identify their own patients once Dyania Health software matches a candidate to trial criteria. The patient’s physician is notified of the match, who can then coordinate with their patient to refer them to a principle investigator (research physician running the trial) and improve enrollment and accessibility to a particular trial.
How does Dyania Health's business work?
Dyania Health provides pharmaceutical companies launching new drugs with a streamlined efficiency that saves them time and money wasted on manual processes for trial enrollment. Each trial Dyania Health contracts requires roughly 1–2 weeks in ramp up to program the specific trial criteria into our algorithms. Afterwards, the pharmaceutical company can implement Dyania Health as a technology vendor at a few or all of their trial sites. Based on the number of trial sites for integration, the company charges a designated base fee in addition to the referral fees for driving a portion of each trial’s enrollment. This referral fee is triggered upon a patient being referred to a principal investigator and varies in amount based on the complexity and rarity of the disease. For example, the ease in finding diabetes type 2 patients varies drastically as compared with pancreatic cancer patients within a specific stage and without having metastasized. Therefore, Dyania Health tailors each contract with a pharmaceutical company to the exact needs and nuances of the particular trial, commonality of the disease, and complexity of protocol requirements.
Where is Dyania Health focused geographically?
Dyania Health’s technology is both HIPAA and GDPR compliant and therefore can be integrated in both European and N. American trial sites as necessary. The company maintains team members in:
• Jersey City, New Jersey, USA
• New York, New York, USA
• Los Angeles, California, USA
• Athens, Greece, E.U.
How does trial site integration work?
Dyania Health sends a team of software engineers to complete integration at each trial site as virtually all trial sites maintain local on-premise servers. The software program is uploaded to a back-up copy of the trial sites patient EMR data base. The software is EMR system agnostic and is specific to the type of database structure (MySQL, SQL, hybrid, etc.). The program runs the anonymization and clinical data extraction once on the entire database and subsequently once per week for incremental changes to the database. Each new trial site to be integrated has an initial set-up timeline of 1 month but once set-up, can be integrated for further trials within 1–2 weeks with additional protocols.
How is the algorithm tailored to specific protocols?
Many protocols have subjective or vague criteria that require interpretation through judgment of principal investigators (physician researchers). Dyania Health works with pharma to clarify what these criteria mean on an objective level, and which medical histories and categories would fit the intention of these criteria. Due to certain limitations of data and medical histories stored in EMR systems, certain criteria might not be referenced in medical histories and records and therefore will not be able to be identified by our algorithms. Also, certain detailed information like full genetic reports may be stored in the system as scanned pdfs of printed faxed reports. Reports stored in this format are sometimes illegible by our algorithms.
We define this ahead of time with the pharma sponsor and the selected trial sites. As a result, we have a clear definition of what qualifies as a near exact match and which criteria have been met by patients identified. We also define an acceptable risk threshold regarding precision and recall where physician notes may be vague, have incorrect information, be missing relevant information, or utilizing uncommon descriptions instead of recognized terminology by the NIH and internationally defined medical libraries.
With regards to adaptations to specific disease areas or therapeutic types, Dyania Health team trains the algorithm with parameters, lexicons, synthetic data, and libraries generated by specialists and principle investigators who have extensive experience performing clinical studies on the particular disease or therapeutic type. This bridges terms common to the specific disease areas’ criteria with common clinical notes data types stored in electronic medical records, increasing accuracy.