I recently attended the Medidata NEXT Executive Forum in New York at the Whitney Museum. Box is in partnership with Medidata Solutions, the leader in cloud-based clinical technology in the life sciences industry. Medidata is building packaged applications on the Box Platform to manage regulated content associated with clinical operations.
The NEXT Executive Forum was an intimate setting that brought together thought leaders, scientists and geneticists to discuss how clinical trials, machine learning, AI and sensor-based technologies are pushing the industry towards more personalized and predictive-based medicine. Notable speakers were John Mattison, Chief Information Medical Officer at Kaiser Permanente, Dr. Robert Green, Professor of Medicine at Harvard Medical School and Sally Okun, VP of Advocacy for Policy & Patient Safety at PatientsLikeMe.
Here were my 3 biggest take aways from the day:
1. More data collection will be the trend of the future with a focus on capturing additional data points over time.
At the event, Sally Okun with PatientsLikeMe, the leading online patient community for people living with chronic conditions presented her company's "Digital Me" initiative. This project focuses on a specific cohort of patients from one of their online communities. They repeatedly go into a patient's home to collect data samples over a longitudinal period of time to help inform how a disease starts, progresses and can be treated. I have heard of similar efforts with other startups in the genomics and behavioral health space.
It seems that everyone in the industry now wants more data and additional data points. This includes data coming from sensors and wearables, remote monitors, DNA for genetic sequencing, or behavioral and environmental signals. New data inputs are now being integrated into traditional clinical and administrative data sets to extend the analysis of disease acceleration and mitigation. A great example of this new trend is Google/Verily's Project Baseline which is partnering with Duke University and Stanford Medicine to track 10,000 patients over 4 years to get a "baseline" for good health to better understand the transition from health to disease.
2. Big Data and AI in healthcare - everyone is doing it!
Working at Box, a leader in cloud content management, we understand the critical role the cloud plays in data storage and analysis. Without the cloud's computational power you cannot sufficiently train machines to do Artificial Intelligence (AI). And because the cloud enables so much more data acquisition and storage, we have seen a dramatic increase in the the number of healthcare related Artificial Intelligence (AI) startups that have received funding this past year. Simply put, more startups and larger corporations (e.g., IBM Watson Health) are applying machine learning to clinical and administrative data sets to:
- Predict the risk of an upcoming negative event in a patient's life;
- Predict length of stay (LOL) in a hospital;
- Improve drug discovery and analysis of potential drug targets;
- Automate machine assisted diagnosis, and;
- Match drugs to a patient's particular genetic make up.
A few interesting examples of how AI is being applied to assisted diagnosis is occurring in the the the imaging space with startups like ZebraMed out of Israel and Enlitic in San Francisco. These companies are acquiring large sets of imaging data (e.g., DICOM images such as X-rays and MRIs) and training machines on how to recognize certain patterns in the images (computer vision) which could ultimately lead to machine assisted diagnosis or regressive quality checks on human diagnosis. Google is also now in this game! They made a ton of news when they bought an AI company called DeepMind in 2014 which is now being use to conduct machine learning on healthcare data from the National Health Service (NHS) in the UK. Google also announced several partnerships through its Google Brain research effort with Stanford Medicine, UC San Francisco and The University of Chicago Medicine to examine EHR data to predict disease.
3. New tools are making clinical trial data capture and recruitment more democratized.
Finally later on in the day, we talked about the state of clinical trials today. Being a leader in clinical trial technology, Medidata had a ton to say about this topic. As a an add on to this conversation, we also talked about all of the new apps that are working to democratize the clinical trial data collection and recruitment process. For example, Apple's Research Kit, which is an open source framework that allows principal investigators, researchers and doctors to build clinical apps that help collect new data points during a clinical trial is now taking off with several academic medical centers building apps for trials they are involved with. Also the startup Medable provides a platform to create secure and compliant mobile apps that leverages Apple’s ResearchKit and HealthKit where no developer is needed. Medable makes it super easy for researchers and doctors to write apps themselves and collect data from mobile phones, wearables and other connected devices to improve data quality and validity. There will be a whole new category of innovations that will work to bring research directly to a patient's home by integrating remote monitoring, smart houses with sensors and telemedicine into the clinical research process. We will no longer be bound by geography for doing research on patients, especially via clinical trial recruitment and patient monitoring.
In closing, the day was full of interesting presentations and trends on data capture and acquisition in healthcare. It gave me great hope that we are closer than we think to having new data sets be the norm for the collective study and prevention of disease. In a sense, each of us in the future may start generating a personal data shadow or real-time, digital health twin.