Health analytics have never had a more visible spotlight moment than right now. COVID-19 is truly testing our ability to harness our health data quickly to derive high-impact insights.

Most of the public-facing analytical attention during this crisis has rightly been directed at understanding the disease’s geographic spread and its corresponding implications for health care providers. Beyond the clinical insights needed to combat this viral threat on the front lines of care delivery, the crisis also provides an ongoing illustration of the data sciences competencies that organizations must develop in order to create and operationalize new insights, including:

  1. Well curated data. Raw data coming out of transactional systems such as electronic medical records systems are a great start, but we need properly curated data – standardized, contextualized, structured for analytics – if we want to speed our time to insight. We can see the payoffs in data governance pretty clearly.
  2. Near-real-time data. It is not enough to copy data periodically; we must become more adept at putting current data at our analytical fingertips. There are too many use cases – biosurveillance, clinical deterioration, etc. – that require we know what is happening with patients right now.
  3. Well-defined use cases. Different needs require different approaches to designing and deploying analytical insights. Consider a few of the current high-impact use cases:
  • Alerting. How do you establish and update signal detection criteria associated with an emerging phenomenon of interest?
  • Situational Awareness. How do you aggregate and manage the key performance indicators associated with successfully navigating complex operations under stress?
  • Case Management. How do you organize, analyze, and surface all of the data needed on an individual case basis in order to assist front-line staff?
  • Disease modelling. How do you define the cohort(s) of interest, and how do you manage them over time?
  • Decision Support. How do you conduct bespoke analyses on an ad hoc basis to support unanticipated questions and dynamic decision making?
  1. Sound, Agile and Advanced MethodsWhen analysis become life critical, it becomes all the more important to focus on well-founded methodological approaches and careful interpretation of results. We’re seeing first hand why taking an iterative, lifecycle-oriented approach to creating insights is a best practice. And we see why descriptive statistics of the past are not enough – we need predictive analytics as well.

As I work with companies to help them develop more advanced, mature competencies around data and analytics, those areas remain the common themes in organizations of all sizes. We have a lot to do.

And yet, as we all navigate this crisis, there are many opportunities to practice gratitude. Beyond family and friends, here are a few I like:

  1. ScienceWe live in an enlightened age that allows us to influence the course of natural events that would otherwise be considerably more tragic.
  2. ScopeWe are learning how to respond to emergent, large-scale and fast-moving epidemiological events on a disease that is not even more virulent and deadly than COVID-19. Though ugly, the current test of our abilities could be much worse.
  3. Innovation. Innovation is not something that “just happens” out of the blue – it is an intentional, skilled act. Consider that we have gone from virtually no awareness of this particular virus strain to having multiple tests and therapies under discovery, development, and use – using both existing and new treatment models – in less than 90 days. They are not perfect, and we will get better, but there is good news in our ability to innovate.
  4. Amazing health care professionals. Unless you work in a health care setting every day, it can easy to forget the endless sacrifices these people make – not just during global pandemics, but every emergency, every holiday, every late-night shift of every day, 24×7, always. To every one of you — thank you very much.
  5. Data. A decade ago, our ability to respond effectively to events like this would have been considerably weaker than we are today. The past ten years of continuous investment in electronic medical records systems, health data interchange, data warehouses, visualization, and advanced analytics have empowered us to move much more quickly than in the past – speed being one of the most important factors in combatting any epidemiological event.

To that end, I’ve enjoyed observing the numerous examples of analytics, data visualization, and data resources that have surfaced over the past few weeks. I’ll provide a start to the list, but feel free to add your own in the comments.

Johns+hopkins+coronavirus+resource+center+screenshot
Johns Hopkins Coronavirus Resource Center. For many people, this site has become their single source of truth about the global spread of the virus.
Covid+actnow.org+screenshot
COVID ActNow.org. This site provides modeling of hospital demand and interventions.
Tableau+covid 19+resource+hub+screenshot
Tableau COVID-19 Resource Hub. Tableau has been engaged to help organizations and the general public understand the epidemic through the use of data visualization.
Sas+coronavirus+report+screenshot
SAS Coronavirus Report. SAS has developed a series of data visualizations that characterize the progression of the global disease.
Image Asset
Tibco Spotfire COVID-19 A Visual Data Science Analysis and Review. This site provides a rather lengthy analysis of data sciences methods in use around COVID-19.
Washington+post+coronavirus+simulation+screenshot
Washington Post Simulation. Harry Stevens at the Washington Post created a very cool simulation demonstrating the effects of social interaction on viral transmission. It sparked my mind with numerous ideas for agent-based simulations.

Github COVID-19 Resource List. A truly massive list of COVID-related resources was compiled by a couple of data sciences at Booz Allen Hamilton (Catherine and Michael); every data scientist working in this space should check it out.

Tags