Contact Us
Navigation
Tweets
Thursday
May242012

The Semantics of Big Data

I had the pleasure of attending the Big Analytics Road Show in Boston this week. The presenters and sponsors did an outstanding job of describing the “big data” ecosystem. They even offered clear descriptions of Hadoop and MapReduce for non-technies, which is quite an achievement.

The most rewarding aspect of the day’s program, however, was its emphasis on how the data can be used to add value to business decisions. Consequently, the focus wasn’t on acquiring massive quantities of data (although zettabytes and yottabytes were mentioned!)—or even on the value of organizing big data sets. Instead, the program provided many examples of how analysis of structured and unstructured data in tandem can lead to new insights that can improve business processes and marketing decisions.

Years ago, at InfoCommerce Group we coined the phrase “data that can do stuff” to describe the advantages of well-designed data products. In essence, a data product that is designed to meet a defined need of a target audience becomes a decision tool when analytics are applied. With the era of big data upon us, even textual data and real-time streams of behavioral data can be leveraged via semantic and pattern matching technologies to obtain data that can do stuff. Furthermore, the different types of data can be overlaid to achieve higher levels of insight into customer behavior or patient outcomes, for example.

The takeaway point: data analysis tools and techniques that used to be available only to big life-science companies and search engines are now entering a phase where the costs make the technologies more widely accessible. However, as someone mentioned at the Big Analytics event, Gartner Group places big data at the peak of inflated expectations on its hype cycle curve. I agree with Gartner because of the level of noise surrounding big data. Nonetheless, with proper alignment between the data, business goals, and execution, opportunities to benefit from big data—or should I say big analytics—exist today.

Sunday
May202012

Between The Lines: Finding the Truth in Medical Literature (Review)

“BTL provides the best critique and comparison of observational vs. interventional (e.g., randomized clinical trials) research studies that I’ve ever read. Even evidence-based medicine experts will find something eye-opening in this book.”

Posting a book review on this blog is a first for me. I am making an exception because this compact volume, Between the Lines: Finding the Truth in Medical Literature, by Marya Zilberberg, MD, MPH, provides an expert’s explanation of many critical issues related to health literacy, evidence-based medicine, and changing models of medical research—all issues that are covered in this blog.

At the highest level, Between the Lines tackles the complex issue of uncertainty in medicine. Dr. Zilberberg presents a framework for assessing the strength of medical evidence in a way that anyone with some basic knowledge of statistics can follow. She uses clear examples that explain, for instance, why a medical test with a 5% rate of false positives could yield a 98% chance of a false positive if the known prevalence of the disease is very low. If these numbers sound irrational, then it’s time you either study Bayesian statistics or read Between the Lines.

In fact, Bayesian statistics are what united Dr. Zilberberg and me. We met via Twitter and our first engaged conversation occurred when she commented on David H. Freedman’s article in The Atlantic: Lies, Damned Lies, and Medical Science. David’s article provoked quite a lot of discussion about the state of evidence-based medicine (EBM), at least based on the type of research we currently consider our ‘gold standard’. His article profiled Dr. John Iaonnidis, who is now chief of the Stanford Prevention Research Center at Stanford Medical School. [1]

When Dr. Zilberberg started explaining the effects of heterogeneity in her blog, I knew I had found someone who had the ability to address important statistical topics in a way that could be understood by a broad universe of readers.

In addition, the book is an excellent resource for non-medical professionals who do have some training in statistics. For me—someone who has experience in econometric modeling and has long been an advocate of Bayesian statistics— but has no formal training in epidemiology, I found the book to be a terrific resource for translating mathematical statistics terminology into medical statistics terminology. All I need now is a self-study guide and comprehension test and I think I’ll feel confident in my understanding of concepts in epidemiology. This shouldn’t be a surprise given that Dr. Zilberberg teaches epidemiology.

I highly recommend this concise volume to anyone involved in peer-review or any aspect of medical communications. I’d even go as far as to say it should be required reading for these groups. And for clinicians and those who determine evidence-based guidelines? Well, I know I’d feel a lot more confident in our healthcare system if I thought that most clinicians could answer the 12 questions that Dr. Zilberberg recommends patients ask before accepting to undergo a medical test or procedure (see Chapter 12).

Finally, I’m confident that Between the Lines will be an important addition to core readings for two groups I highly admire: 1) medical librarians and 2) the Society for Participatory Medicine (http://participatorymedicine.org/).

To obtain a copy of the book, which was published May, 2012, visit the Between the Lines website.

 


[1] See: http://alumni.stanford.edu/get/page/magazine/article/?article_id=53345 for a recent article about Dr. Iaonnidis’s work.

Tuesday
Apr032012

Secondary Data Usage in Healthcare

I was guest speaker at the March 22, 2012 “Let’s Talk HIT” series hosted by Scratch Marketing & Media in Cambridge, MA. The topic I chose was Secondary Data Publishing in Health. Health Content Advisor’s parent company, InfoCommerce Group, has a long history of guiding business media companies in constructing data products, but increasingly we are finding interesting examples of secondary data products that develop as a by-product of technology companies. Electronic Health Records (EHRs) represents one of the more compelling examples of information technology that has the potential to spawn a new generation of data products.

Scratch Marketing has posted the video of the talk, which was structured as an interactive group discussion, in 8 parts. See their YouTube page for the list of segments: http://bit.ly/H9Wjk9.

See the event recap by Lizzie McQuillan at Scratch Marketing here:

http://scratchmm.com/2012/03/event-recap-let%E2%80%99s-talk-hit-with-janice-mccallum/

Also, for a provocative view, read Marya Zilberberg, MD, MPH’s takeaway from the evening’s discussion:

http://evimedgroup.blogspot.com/2012/03/how-our-healthcare-spending-is-like.html

Thanks again to Scratch and the many Boston-area (stretching all the way out to the Berkshires!) health IT, public health, healthcare publishing, entrepreneurs, and marketing experts who attended and participated in the discussion. Scratch Marketing added Twitter handles to the video, which helps tremendously in identifying each speaker.

Monday
Mar192012

Better Security Solutions Build Trust

The importance of building and maintaining trust among entities that are collaborating or exchanging goods & services is paramount. Without trust, markets where imperfect or asymmetrical information exists won’t operate efficiently. Whether we like it or not, “imperfect” and “asymmetrical” define the current mechanisms used to exchange personal health information.

At the same time, new levels of collaboration are being introduced between providers, payers, researchers, and patients within tightly or loosely-defined networks that participate in the care of a patient. All of this collaboration requires multi-directional exchange of data. In a guest commentary post on Health Data Management Magazine, I emphasize the importance of employing appropriate security solutions to help build and maintain trust in the integrity of data partners.

To accompany the post, below are links to recent news on the companies I mention in the article:

Amcom Software Enhances Physician Communication With Release of Amcom Care Connect

http://www.amcomsoftware.com/News/02-28-12.aspx

Imprivata Launches Secure Texting App

http://www.imprivata.com/company/press/imprivata-launches-secure-texting-app

PerfectServe Secures $10.9 Million in New Financing

http://perfectserve.com/aboutus/releases/release020212.html

Samsung Selects Voalté Clinical Communications Solution of Choice

http://voalte.com/NewsPressDetail.aspx?Id=56&title=Samsung-Selects-Voalt233-Clinical-Communications-Solution-of-Choice

TigerText Secures More than $8 Million in New Funding

http://www.tigertext.com/tigertext-secures-more-than-8-million-in-new-funding-adds-veteran-life-sciences-and-technology-investors-to-its-board/

Vocera Introduces Vocera Connect for Leading SmartPhones and Cisco Wireless IP Phones

http://www.vocera.com/index.php/company/pressreleases/120215

 

See also this video interview with Perficient from HIMSS: http://www.healthcontentadvisors.com/blog/2012/2/28/video-interview-from-himss12.html

 

Tuesday
Feb282012

Video Interview from HIMSS12

For those looking for the Women in HealthIT video, see:

http://t.co/pfbTB59K.

What has now become an annual event: an interview between Liza Sisler from Perficient and me at HIMSS. This year, I talk about 1the effect of Accountable Care Organizations (ACOs) and value-based purchasing in general on the healthcare industry and the underlying theme of trust and security in health information exchange.

http://blogs.perficient.com/healthcare/blog/2012/02/27/himss-2012-interview-with-janice-mccallum-phi-and-patient-trust-video/

I’ll post more on the topic of trust and security of personal health information (PHI) soon.