At this year’s HIMSS conference the big buzz was analytics. Seemed as if every vendor in attendance was promoting some sort of analytics capability. It was a tad overwhelming and if I was having a problem trying to separate out all these countless vendor offerings (this is exactly what I am trained to do as an industry analyst) I couldn’t imagine what it must have been like for those representatives of healthcare IT departments.
Then I had this deja vu moment – I had experienced this all before a few years earlier when the buzz at HIMSS was health information exchange (HIE). As with the HIE market back then, it was hard to make sense as to who was riding the latest marketing buzzword (e.g., ACO analytics) and who was actually solving problems for customers. There was also no easy way to differentiate one solution from the next. Clearly, the market needed someone to make order from the chaos.
Making order, applying structure to a market is one of the key roles of an analyst firm. As Chilmark Research has done for the HIE market, we intend to develop a similar structural framework for the analytics market.
We have recently embarked on a major research project to delve into the healthcare analytics market that will culminate in a report of similar length (~100pgs) and breadth to our 2012 HIE Market Report. Padmaja Raman, who recently joined Chilmark Research, will be the lead analyst for this research project.
In advance of formal research, we have had roughly a dozen casual conversations with various stakeholders in the healthcare analytics market. Here’s a few things we have learned so far:
Two dominant market vectors: 1) Changes in reimbursement and associated risk. 2) Regulatory requirements (MU attestation, quality reporting etc.)
Analytics using clinical data is in its infancy.
Clinical datasets are typically a mess and very difficult to work with.
For foreseeable future, most healthcare organizations will need to rely on claims data to ascertain risk.
We may be years away from doing true real-time analytics with clinical data.
Most healthcare providers do not have well-trained informatists on staff (particularly problematic for smaller healthcare organizations) and thus “analytic offerings” will have a high service component.
Business intelligence (BI) tools are not well-developed (either too simplistic or too complex for average clinician) and thus clinicians are unable to run their own reports. This has led to IT departments being completely overwhelmed with clinicians asking for various reports.
Those pretty dashboards that are based on clinical data which you see at trade shows and in PowerPoint presentations can rarely be replicated at a client site – lots of smoke and mirrors.
Most EHR vendors are just getting started in this area and their solutions are rudimentary.
This is a fascinating market and one where I am confident we can provide value. Without exception, every stakeholder we have spoken to has stated that this market is prime for in-depth analysis. Now it is our turn to provide such – Stay Tuned.
You nailed it with this summary! I can only imagine how thorough a 100 page report would be.
How can there be BIG data or analytics if there is NO data?
As a first step hospitals have to start collecting clinical data. By simply implementing an EHR this will not be accomplished. Most hospitals still don’t understand this and they believe that a certified EHR will solve the problem.
The biggest challenge will be for the hospitals to get the clinicians to enter the data as discretely as possible. Here underlies the strongest barrier: “hospitalists cultural behavior”. Those of us that have been in healthcare IT for a long time know that no matter how much money you pour into technology, if you don’t change the attitudes of the users, progress is null.
Some technologists are offering NLP (Natural Language Processing) as the panacea to solve the clinical dataset mess. I’ve been in several of these projects and this is not easy. Even Mr. Watson can’t figure out most clinical statements.
Using claims data for clinical analytics is never going to work. Claims data is for financial and possibly some operational analytics. Research and quality also demand precise data.
Achieving real-time analytics would place the technology companies and the hospitals in the medical device manufacturing arena. Many companies that are offering BI tools have never been in the medical device vertical and have no idea how onerous it could be to work in such a regulated space. BI and real-time analytics would require hospitals to put in place custom software development shops. Are hospitals ready to put in place a rigorous quality control system? If not, I can easily predict a spike in litigations regarding patient harm.
All this shouldn’t be a discouragement for hospitals to get started. The problem is that they get easily trapped by “analysis paralysis”.