Time-to-insight… It’s a metric you don’t hear about much. But it’s a lurking factor throttling the rate of scientific progress in cancer research. So what’s holding us back?
Simple analytic questions can’t be answered easily today. That may surprise many people, but not clinical researchers and their organizations. For example, questions like “Which patients with a pathogenic mutation in the KRAS gene had an exceptional drug response” can take a months-long project, rather than seconds or minutes, to get an answer. That slow-walk to building up insight is repeated every time a question is asked, no matter how large or small the research organization making the inquiry.
The amount of work required to go from real-world clinical data to research-ready inquiry is far more than most people know. In the clinical environment where most biomedical data exists, the primary digital representation of a person is an electronic medical record (EMR) made up of a series of encounters, supplemented by vendor archives for molecular, genetic sequencing, imaging, and other data. While EMRs are useful for clinical care, when a patient is in front of a physician, a medical record isn’t a useful model of a patient for research purposes. Historically, answering research questions has first required the laborious work of aggregating, harmonizing, and curating this multi-modal data. The trouble is, that process has been largely manual and resource intensive. It’s the main limiter of research time-to-insight and the driver of long wait times for fundamental questions.
The research community has been missing the right technology infrastructure. They incur substantial costs in manual steps and manual data wrangling. They are limited in their ability to include more patient data modalities, to follow patients over very long periods with more frequent data, and to be more agile in serving new use cases. Research is often stuck with decades old survey techniques and an infrastructure that exacerbates data silos. This gap in infrastructure is a drag on the cycle of research and practice across medicine.
Our vision at Manifold has been to build an end-to-end AI-powered clinical research platform for the new kinds of research studies that will drive evidence generation in the coming decade. What does end-to-end mean? It means to make it easier for research teams to: enroll and engage study participants; get consent and retrieve medical records; collect samples; collect patient reported outcomes; aggregate data from internal and external sources; harmonize and curate longitudinal data; find and analyze multimodal datasets, and collaborate securely across organizations.
We’re grateful to our design partners like Indiana University Melvin and Bren Simon Comprehensive Cancer Center and Winship Cancer Institute of Emory University whose constant and valuable product feedback has enabled us to design a product that will accelerate time-to-insight and increase efficiency for all of our customers. We’re also grateful to an ecosystem of technology partners, including some of the biggest infrastructure companies, who have enabled us to advance our ambitious roadmap faster than anyone thought possible and most importantly to serve our customers better.
I have always believed in the African proverb, “If you want to go fast, go alone. If you want to go far, go together.” At Manifold, we are convening a community so passionate about addressing these challenges that we will be able to go farther together than ever before.