myTomorrows Launches “Treatment Search” — A Search Tool to Help Physicians Find Clinical Trials
myTomorrows, a platform that helps patients with an unmet medical need discover and access treatments in development worldwide, has unveiled the myTomorrows’ Treatment Search. The Treatment Search is a proprietarily developed search tool that gives physicians and patients the ability to rapidly perform a global search of registries to find Clinical Trials and identify relevant pre-approval treatment options. This AI-powered tool consists of an interface that is intuitive, presenting search results in a manner that is easily understandable, readily usable, and backed by our extensive user-research testing with physicians. Treatment Search is publicly available and free of charge for physicians and their patients.
The limitations of Clinical Trial registries
The number of Clinical Trials is increasing dramatically in response to the growing burden of disease around the world, but there is still a need for clinical research to be more inclusive. With the number of Clinical Trials on the rise and the global availability of information, new solution-oriented platforms to find Clinical Trials are being envisioned by public and private entities. myTomorrows has taken an early lead in creating a tool to address a multitude of challenges when it comes to searching for and accessing pre-approval treatment options. Our business model strives to position the patient (and their physicians) more centrally in the drug development process.
This increased availability of information and the rise in patient empowerment has drawn attention to the challenges that physicians (and their patients) experience when using public registries to search for pre-approval treatment options (e.g. Clinical Trials and Expanded Access Programs). Clinicaltrials.gov lists hundreds of thousands of trials from all over the world. However, the reason for the existence of this registry and others, is for researchers to list all information about all trials — as is required by law — and not necessarily to provide physicians and patients with information about suitable treatment options. As a result, these government registries can end up being quite cumbersome and not best suited to physicians and patients trying to find suitable Clinical Trials.
Current tools (mainly clinical registries) that physicians/patients may use to find Clinical Trials have at least several of these limitations:
- They do not have global coverage
- The data is not necessarily up-to-date, despite regulations and corresponding data quality efforts
- Search performance (accuracy of retrieved trials, retrieving all trials that are relevant) could be better or may be unmeasured/unreported
- A design that is not necessarily tailored to treatment search as task
As all of the above are needed for a good search performance and experience, we work towards applying existing methods and inventing new ways to make the system better in all of these aspects.
Traditional registries may include irrelevant studies for a search query (trials for the wrong condition searched for, or trials that are no longer active) or fail to prominently display patient or physician relevant information along the lines of geographical location, disease type and recruiting status. They are not fit for the purpose of helping physicians and their patients find Clinical Trials because users must first scroll through all the information provided in a study record to find the information and decide on their eligibility for the trial. These challenges contribute to the fact that many patients never find Clinical Trials that may be available to them, let alone consider participating in one.
Bringing treatment options to patients
Having information spread across multiple Clinical Trial websites may be incomplete, inconvenient and inefficient for physicians. Registries cover what they are mandated to include. Treatment Search solves that by accurately and rapidly identifying trials and programs in various databases around the world that are: actively recruiting, highly relevant and most likely to be suitable for your patient.
Today a physician using Clinicaltrials.gov or EudraCT to find Clinical Trials and identify treatment options can search through over 350,000 and 38,000 respective research studies being conducted in more than 200 countries. Our search tool searches CT.gov and EudraCT at once, reviewing about 390,000 trials and returns a more comprehensive and redesigned display of results with a particular focus on UX design — in one search. Some of the trials in our database (about 19,000) are registered in multiple databases and our technology recognizes them as duplicates, so our database contains about 370,000 unique trial listings. This allows for high coverage with superior search-quality and performance when conducting a search.
In short, myTomorrows has tried to address:
- Data coverage (Data from multiple registries)
- Data integrity (Duplicates, status and dates)
- Search performance (document to query matching)
- Search experience (UX)
As the limitations of these government-mandated registries have become more apparent, a number of trial navigator websites have emerged. These websites address the qualitative issues facing physicians and patients who are trying to find Clinical Trials by way of providing better interfaces and arranging information in a more user-friendly fashion. However, the majority of these sites are siloed, reliant on the clinicaltrials.gov database, and don’t offer a comprehensive solution (including patient support and navigation) to finding the right trials for a patient. Many of these websites are understandably restricted to their own specific interests such as research, disease (advocacy), or corporate.
Comparison of Treatment Search to leading clinical trial databases
A more comprehensive way to find Clinical Trials
We listened extensively to physicians during many months of user research. As a result of what we learned, we built our Treatment Search with more sophisticated search options for physicians. We also created a better experience by reframing the existing functionalities and altering the categorization and default filters when people interact with the search. (see numbers 2 and 3 in figure below)
We paid specific attention to increasing the performance of finding relevant trials. More specifically, it is known that public registries that are used to find Clinical Trials often present non-recruiting and/or less relevant research studies. Treatment Search identifies active clinical research studies to display as the default results for a search query. (see number 3 in figure below)
Treatment Search query display
Treatment Search aims to be more comprehensive than Clinical Trial registries, while also about displaying data and offering support tools that aid physicians in their decision-making process.
An option we offer for physicians and patients is the Treatment Search Report. Our medical team can be contacted directly via the search platform to help further narrow a search based on each patient’s medical records, including the genetic profile of a tumor or treatment history to create a personalized list of possible treatment options for a specific patient. Our Patient Navigators are available to guide patients and their physicians through the process of getting a Treatment Search Report. However, we appreciate that not every physician needs a report to be compiled and may want to do a search on their own. By making Treatment Search available to physicians, we can help them make their decisions more quickly and offer their patients every treatment option that may be a possibility for them.
User experience with physicians led to an intuitive design
Our development process includes extensive user research with physicians in order to fully understand the physician’s workflow and efforts to help patients. Treatment Search illustrates how we understand the physician’s decision-making process and continuously make improvements to the platform that supports them on their journey to access a pre-approval treatment for their patient.
myTomorrows utilizes key principles of human-centered design to ensure that the interface is highly user-friendly, intuitive, and optimized for use across different devices. Additionally, the results are presented and categorized in an intelligent, easily scannable way that reduces information overload.
Using technology and design to re-envision how physicians find Clinical Trials
While websites like clinicaltrials.gov have undertaken modernization efforts to improve the user experience and better help patients search for Clinical Trials, we believe that we have ultimately created a more comprehensive and user-friendly way to find Clinical Trials.
In the past decade, the field of AI has gone through rapid development, that has seen more accurate — and implementation-wise more mature — AI-based products being developed. The medical field has experienced a similar trend, with various data modalities such as images and text being processed with new AI methods. Our search makes use of these advances, relying on state-of-the-art Machine Learning (ML), Knowledge Representation, and Natural Language Processing (NLP) methods. We use these to understand medical text, which is critical to ensuring that searchers find the right treatments described in clinical trial, EAP, and patient-related documents.
Being able to process and understand diverse document sets (e.g. data from various registries), having state-of-the-art search algorithms, as well as interaction design being applied, we are on course to continuously deliver a search experience that is associated with high data coverage, high search accuracy, and high user engagement.
A knowledge graph allows for several search functionalities
The basis of our search engine is a knowledge graph we are constantly developing to capture what we may know about conditions, genes, interventions, possible trial locations, sponsors, as well as relationships between these. With this knowledge, we can understand medical texts better, and improve our search algorithm. The knowledge is compiled from various structured data sources such as the Unified Medical Language System (UMLS). We aim to consolidate this knowledge to create an as complete and accurate knowledge graph as possible. This knowledge graph is the basis for several search functionalities (query-trial matching, query auto-completion), and can also be the basis of other functionalities, as well as products that may serve people with unmet medical needs.
The design gives us a rich knowledge base derived from medical conditions and key variables, and it allows us to understand and search complex medical terminology and documentation. The way we can understand Clinical Trial documents is similar to how we can understand other sources of information, such as patient files and publications.
The in-house proprietarily developed technology gives us the flexibility to build on the existing knowledge base and further address certain stakeholder needs. Our approach offers many distinct advantages over traditional registries.
By default, our technology set-up enables us to custom-build applications for a wide range of use cases. We can relatively easily add other important variables such as genetic mutations, intervention, sponsors, and geo-locations. These functionalities are implemented based on technical expertise and tailored to the specific requirements of our stakeholders.
Beyond its superior search capabilities, Treatment Search displays data in a way that supports the physician in their decision-making process. Treatment Search displays results in a more user-friendly way that is comprehensive yet concise.
Listing PubMed Abstracts
In addition to presenting listings of Clinical Trials and Expanded Access Programs, our Treatment Search displays a list of related articles from scientific publications and medical journals from PubMed that may support physicians in their analysis of the treatment options listed.
These lists are based on both analyzing links from Trial data to PubMed and extracting links/references from PubMed to trials, Thus, the treatment Search includes 600,000 PubMed links, and presents these articles with abstracts so physicians can review the key findings from these studies without having to switch to PubMed and can stay focused on the task at hand of deciding on suitable treatment options for their patient.
Study-linked publications
Establishing equity in the availability of information and access to drugs requires a complete redesign of the treatment paradigm with a renewed effort to bridge the gap between clinical research and care. Our Treatment Search lays the foundation to begin to realize this goal.
In a world where the traditional pharma distribution model is looking for new options, myTomorrows has taken a leap forward by launching the Treatment Search, which is based on a knowledge graph, which can be used to build other applications relying on that technology. Insights into what patients and physicians are searching for online could facilitate a more patient-centric drug development model. Solving the problem of Clinical Trial participation is yet another key focus of myTomorrows’ integrated platform.
Our Treatment Search is just one way that myTomorrows is making drug development serve everyone better. We believe that patients and their physicians — regardless of where they live — should have information about and access to all possible treatment options around the world. That is why we offer Treatment Search and the Treatment Search Report as a free service for patients and physicians.
Learn more about our services for physicians and try the Treatment Search.
*** Special thanks to Zoltán Szlávik, AI Research Director at myTomorrows, for his generous assistance with writing this blog, and to the entire Tech team at myTomorrows, led by Chief AI Officer Robert-Jan Sips, for developing the Treatment Search.
