Lessons in Healthcare and Big Data from HIMSS 2015

Categories: General

HIMSS 2015 exceeded its promise of returning attendees to “their care settings to play an active, transformational role for health IT in their communities” and pulled the largest participant count in its history.

Our education session on “Better Care with Big Data” drew one of the largest audiences of the week as we joined leaders from Kaiser Permanente, Cerner, and Children’s Healthcare of Atlanta for a roundtable discussion about the state of the art in information-driven healthcare with Hadoop. The conversation ran from eradicating hospital-borne sepsis to the need for better metadata to navigate and enrich EHRs and from improving NICU outcomes by reducing passive sources of infant stress to striking the right balance between data size and quality in bioinformatics.

As I roamed the HIMSS expo and met with clients and partners, three themes emerged most prominently from this year’s event:

  • Continuing momentum on precision medicine as an approach to care
  • Consistent focus on quality
  • Amazing progress on devices with an eye toward their interoperability and the data they generate

According to NIH, precision medicine is “an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person.” As precision medicine evolves, it will more and more squarely fall into the purview of omics. Genomics, by way of next-generation sequencing, is emerging as a key practice in research hospitals and will soon be prevalent in other clinical settings. However, other omics, such as proteomics, metabonomics, and exposomics, will accrue their fair share of research dollars in the future. The exposomics funded by US National Institute for Environmental Health Sciences is just the beginning, with great hopes for the HERCULES project at Emory University, the HELIX project in Europe, and the Human Exposome Project. The more work that is done on epigenetics and establishing phenotypic effects of all environmental triggers, the closer we are to finding substrates and biomarkers that will be the key to mediating negative effects of individuals’ precise environments.

Where is the big data in omics and epigenetics? Next-generation sequencing and the trend toward whole genome is making petabyte a much more common term today outside of the historic bastions of bulk data like The Broad Institute or Wellcome Trust Sanger Institute. Today, tens of thousands of samples are more common, and using high performance or storage computing to do math at this scale can be somewhere between optimistic and painful. The continuous and often non-quantitative nature of environmental data makes it have even fewer plausible storage opportunities outside of Hadoop. At the end of the day, these trends are good news for those embracing the notion that future research funding will be most attainable by those employing a scalable big data platform.

HIMSS had excellent examples of the search for patient safety, showing responsiveness to recent CMS orders on hospital-acquired infections. While predicting sepsis and readmissions certainly dominated the dialogue, one could see evidence of wider efforts and forward thinking now that multiple predictive options exist for the clinic. Safety presentations included “Would You Bet Your Mother’s Life on the Safety of Your EHR?,” “Ten for Patient Safety,” and “Alarm Fatigue: We Have It!”

Efforts for predictive medicine around patient safety are in their infancy. The industry will leave the current Pavlovian mode of governments creating penalties for specific events and the industry reacting and investing. How? The first step is to create prediction platforms that are not statistical software, not data scientists, and not checklists, but are methods for applying multiple algorithms against the same broad and deep data that work with all tools, can be run by a data analyst, and can result in remediation and action customized by the healthcare organization.

It seemed the number of new and innovative devices on display this year at HIMSS outstripped anything from the past. However, this year, there was more focus on the information architectures behind the devices as attendees were more serious about implementation at scale. For example, the sponsors of the Open-Source Integrated Clinical Environment (OpenICE) were actively working the show, sharing their message that the more devices in clinical settings, the more important the mission to make their data work together.

Whether the focus was devices, prediction, precision, safety, or outcomes, every discussion eventually turned to the fact that this new approach will be generating much more data than has been generated before, or will need to analyze much more data than has been analyzed before.


2 responses on “Lessons in Healthcare and Big Data from HIMSS 2015

Leave a Reply