AiCure
The Next Frontier in Health Innovation

During a clinical trial, accurate assessments of a patient’s response to treatment are critical to measure the state of their disease and a drug’s impact. However, traditional measures of a patient’s wellbeing, including sporadic in-person assessments or self-reported outcomes, only capture a single moment in time for that individual. The stretches between these check-ins can tell an entirely different story, filled with potential insights about a person’s condition, response to therapy, or degree of engagement in the clinical trial.

AI-powered digital biomarkers deployed on phone-based platforms may allow sponsors to remotely capture these individual patient responses – from changes in facial characteristics, vocal patterns, movement and more – offering a unique and objective view of disease progression over time. Despite this potential, the proprietary nature of many digital biomarker algorithms renders them inaccessible to the scientific community, hindering their validation and improvement. AiCure’s open-source platform helps to break down these barriers, foster collaboration, and allow trust for novel measurements to be built in the public domain. Democratizing access to these algorithms and welcoming diverse perspectives across the industry to contribute to their development can open a world of possibilities in understanding the nuances of patient behavior, and driving equitable, personalized care.

The power of digital biomarkers

The potential impact of digital biomarkers can best be explained through an example. Imagine someone who has recently been diagnosed with depression and has started therapy with any of a number of available medications. For several weeks on initiation of therapy, they may start experiencing effects of the medication such as sedation that negatively impact their quality of life. When they visit their doctor for a check-up, they may find it difficult to disentangle these effects from the symptoms of their disease. This is not only a highly frustrating experience for patients, but can negatively impact the ability of doctors to efficiently optimize an individual’s medication and dosing.

Instead of relying on subjective self-reported perceptions of how a patient is feeling, digital measures collected on phones can treat these assessments more like an engineering problem, measuring a patient’s response using computer vision algorithms applied to video and audio collected remotely. Especially for conditions with symptoms that have visual and auditory characteristics such as schizophrenia or Parkinson’s, an objective, consistent way to track a patient’s response can help elevate a trial’s data and accelerate research. Through AI-powered analysis of patient video and audio data, clinicians and researchers can pinpoint these crucial disease characteristics to personalize a patient’s care and better understand a patient’s lived experience with their illness.

While the future of this innovation is exciting, it’s often under lock and key as proprietary technology. The reality is that the quality development of these solutions is an endeavor that no one company can take on alone. Open-source AI platforms that bring together diverse industry perspectives and data are vital to making digital biomarker technology a part of everyday research and patient care.

Open-Source platforms: explore, expand, and accelerate innovation

Peer review is just as important for AI platforms as it is for any scientific endeavor. In the same way that a cybersecurity company may challenge hackers to crack their code in an effort to make it more secure, open-source AI platforms allow the scientific community to critique, validate and contribute to algorithms. AiCure’s OpenDBM allows researchers to apply AiCure’s digital biomarker algorithms to their own datasets inside their institutional firewalls, letting them ‘roll up their sleeves’ and become familiar with the tools for understanding patient behavior with these unique data sets. While it may seem odd for a company to lift the veil on how their technology works, this transparency is necessary to build widespread confidence in the value these novel assessments provide in understanding disease and patient behavior.

Open-source AI platforms that bring together diverse industry perspectives and data are vital to making digital biomarker technology a part of everyday research and patient care

The scientific community has a responsibility to ensure AI tools are developed in the right way and serve all of our patients. Making sure algorithms are built on an adequately diverse data foundation so they work effectively and accurately with broad patient populations takes more time and resources than most companies can spend themselves. Open-source AI platforms open the door to a variety of industry voices and perspectives to explore, expand and accelerate the diversity of these novel tools. Interrogating algorithms through an open-science framework not only helps algorithms learn, but also determine their performance with diverse patient populations and disease states.

Delivering on the promise of AI

By allowing us to tap into a patient’s everyday experience with their treatment, digital biomarkers could ultimately transform how we draw conclusions about a drug’s impact and its value in a real-world setting. But, there’s work to be done to open more eyes to the value these technologies can provide, and to ensure that they are built to deliver on their promise. Creating a community that encourages frank commentary on the reliability and diversity of these methods is key to securing a future for digital patient assessments.