Getting Juice From Health Analytics


"Is the juice worth the squeeze?"

That's how a colleague of mine asks if we are likely to uncover significant insights from a proposed project in our health innovation center. The question gets right at the heart of executive mindshare regarding health analytics investments; I just wish I had an easier answer.

(Source: Keith Williamson
via Flickr)
(Source: Keith Williamson
via Flickr)

If an organization is not doing any analytics, the answer is always "yes." Anyone arguing that information-based actions are not necessarily better than uninformed actions should be challenged to a blindfolded game of darts.

But comparative valuation can be tough. Before the introduction of software into the everyday workplace, productivity improvements looked easier. If an office was using typewriters, ledger books, and postal mail to conduct its business, the introduction of an electronic word processor, spreadsheet, and email had obvious, measurable impacts in speed, accuracy, and efficiency. But once an organization moves beyond that technologically barren phase, justifications for a different word processor or email platform become more difficult. "Why is it better than what we have?"

When that answer seems challenging, it may appear that there must be "less juice," but that conclusion is often unjustified. Automation, for example, is often a step-wise progression from an existing technological base (e.g., compare Mint.com to older versions of Quicken). In many cases, exponential value innovation requires the presence of existing technologies -- consider, for example, genomics.

Healthcare is just now emerging from its own technologically barren phase of paper records. Similar to the massive enterprise resourcing planning (ERP) system implementations of the past, healthcare will eventually learn that the value of information resides not in how you collect it, but in how you consume it.

But what if you already have basic business intelligence (BI) in place... how can you tell whether "advanced analytics" (e.g., predictive modeling, simulation, data mining, optimization, etc.) are better than plain old counts and averages? I always try to answer truthfully that we won't really know the answer until we do the work. But would you "bet your farm" on the Farmers' Almanac, or do meteorological models and weather forecasting help maximize crop yields? Would you blindly loan a stranger large sums of money, or would a credit-risk score help protect your money?

If there is value in knowing...

  • How likely it is that a particular thing will happen -- where, when, and to whom?
  • Why is it happening?
  • How to get the best outcome when dealing with it?

...then there is "juice" in advanced analytics. And the value is at least equal to the cost of whatever bad thing(s) happen when you cannot answer those questions.

Of course, my friend's question also highlights how hard we currently have to "squeeze" to get insights. All too often, our analytical opportunities are compromised by immature enterprise architectures and information management practices. It can be time consuming and costly to get any insights, let alone insights that are useful.

But no one would have ever designed a "juicer" machine until they got tired of squeezing produce by hand. In my experience, analytical leaders find ways to serve two goals simultaneously: answering the immediate questions and iteratively building a platform for future insights. And that platform becomes a gift that keeps on giving.

What do you think? Is the juice worth the squeeze?

Editor's note: Jason will be joining us for a live, on-air interview this coming Monday, Oct. 7, at 2:00 p.m. ET for a discussion on "Making Medicine Smarter." Register now and tune in next week!

Jason Burke, Senior Advisor for Advanced Analytics & Innovation, UNC School of Medicine

Jason Burke is senior advisor for advanced analytics and innovation at the UNC School of Medicine andUNC Health Care System's Center for Innovation. He currently serves on the leadership team for theNational Collaborative for Bio-Preparedness, and is a principal at Burke Advisory Group. With a career spanning clinical research, organizational development, and enterprise information technology, his current work focuses on health industry transformation through strategy design, data sciences, and emerging technologies. His most recent book is Health Analytics: Gaining the Insights to Transform Health Care (Wiley, 2013). Previously, Burke was the founder, managing director, and chief strategist for the SAS Center for Health Analytics and Insights, an industry-leading think tank pioneering novel approaches to health informatics. He also served as SAS's global head of health and life sciences technology research and development, as well as founder of the $3 billion software firm's health and life sciences global practice. He has held strategy and management roles in organizations such as Microsoft, Quintiles Transnational, and GlaxoSmithKline. He is a cognitive neuroscientist by training, holding degrees from Virginia Tech and the University of Missouri at Columbia. His work has appeared in numerous publications including Forbes, InformationWeek, Health Management Technology, PharmaVoice, Drug Discovery and Development, and BioIT World.

Patient-Centered Data-Driven Care: Carolina Advanced Health

Carolina Advanced Health is a model for a patient-centered medicine that places a heavy emphasis on team-based care.

Courting Better Health: Time to Focus on Health Analytics

No matter your political views, the Supreme Court ruling upholding the Affordable Care Act has removed much of the uncertainty that has dogged the healthcare industry the last two years.


Juice analytics :)
  • 10/21/2013 10:50:49 PM
NO RATINGS

It all has to do with the quality and consistency of the data going in to make a perfect juice. Incorrect data will yield results that may show up in the wrong juice being produced. 

Re: Welcome!
  • 10/7/2013 12:54:04 PM
NO RATINGS

Great article Jason!  I'll never look at fruit the same again now.

Re: third last paragraph
  • 10/6/2013 10:52:18 AM
NO RATINGS

Indeed analytics is good overally but then again the specifics of the fruit to be squeezed also make the difference. Even so, i hope that having a good analytics/BI system will allow the analyst to sieve out the useful from the useless at a reasonable cost.

Re: Welcome!
  • 10/5/2013 12:23:31 PM
NO RATINGS

Thanks Beth, glad to be back and looking forward to Monday's radio show!  Agree with all of the wonderful comments on the post.  The discussion on the nature of the fruit I think is very appropriate.  The current focus on retrospective "quality metrics" is a case in point -- they provide value (as many descriptive statistics do), but in many ways they inhibit asking more sophisticated, powerful questions (which need to cover basic research, clinical trials, and healthcare delivery systems).  So much opportunity. :)

Re: Squeezing Juice
  • 10/4/2013 11:39:35 AM
NO RATINGS

I agree with your comment Nick.  We bring lots of foggy information to the fore with Clinical Trials and other forms of medical research.  Yet it is accepted as evidence.  There is no silver bulit.  i do think that Analytics is a good investment and better than many forms of Medical Research. 

Squeezing The Fruit
  • 10/4/2013 9:44:42 AM
NO RATINGS

Yep, getting the juice is the best way to go. But, make sure the right questions are asked and the data is relevant to the solutions. If you want orange juice don't try squeezing apples. And if orange juice seems to come out of apples, there's something grossly wrong with the process.

Re: third last paragraph
  • 10/4/2013 7:01:57 AM
NO RATINGS

Building that system to actually produce some productive juice is going to be a problem. If it can produce an even better fruit with a bigger yield then it is great. But what if the whole fruit was made with the wrong genetic engineering model? In other words have they asked the right questions and used the right fertilizer?

Re: Welcome!
  • 10/3/2013 4:14:38 PM
NO RATINGS

I agree. It all has to do with the quality and consistency of the data going in. If we can work with good data, we can make good jiuce. Bad data will yiled something else.

third last paragraph
  • 10/3/2013 12:43:00 PM
NO RATINGS

That is where you find the 'stuff' I was thinking of when I read this.  Some systems, and I've seen dozens, are just poorly designed.  The squeeze becomes growing a new piece of fruit. (I love this squeezy juicy analogy).   

Then you present your new juice and and know one knows what it is.  What about that?  Maybe I've taken this too far.  I guess if you re-built the system properly, you just grab a new piece of fruit right?  for just a little squeezing.

Re: Welcome!
  • 10/3/2013 12:41:41 PM
NO RATINGS

Jeff, right, it is in a lot of ways one big, stinky, sticky mess. That's one of the reasons I'm really looking forward to having Jason on air with us Monday, for his insider's perspective on how to make it better.

Page 1 / 2   >   >>
INFORMATION RESOURCES
ANALYTICS IN ACTION
CARTERTOONS
VIEW ALL +
QUICK POLL
VIEW ALL +