Apixio is one of the few startups tackling this issue. Their analytics engine indexes the underlying data, processes queries and provides context-relevant results. The core technology is supposedly based on Apache’s Pig (a data-flow language and execution framework for parallel computation), Hadoop (a framework that allows for the distributed processing of large data sets across clusters of computers) and Cassandra (a scalable multi-master database).

There are a number of terminologies (read ontologies) in healthcare, trying to specify the concepts and relationships from a particular perspective. LOINC, ICD, SNOMED, CPT are common examples, but see a pretty comprehensive list of all human-related ontologies at BioPortal (filter by category ‘Health’).

So a medical-grade search service offering would need to traverse such terminologies and surface all relevant, normalized data related to the query. For example, a search for keyword “breathlessness” in a patient with long, complicated medical history would bring up documents and encounters that mention items like wheezing, PEFR, smoking, asthma management. It’s no short order to do all that analytical crunching.

Sophisticated data transformation and abstraction offerings are certainly needed for making sense of complex healthcare data. Niche efforts like Apixio, 360Fresh, are signs of growing market realization that the era of just trying to digitize healthcare data is getting over. Now we start figuring out what the heck to do with all the incoming bytes.

PS: Advanced analytics offerings in healthcare are an interesting topic. See this wiki page for a living list of relevant companies in this space.

Believe it or not, a lot of electronic medical record archives today consists of documents in free text format- no structure or organization, just vanilla narrative text. 360Fresh uses their proprietary data-mining logic to extract meaning from that. Generally speaking, I think there is potential for such offerings; especially when presented in a focused manner. For example, a service that identifies high-risk patients in ED or Labor & Delivery patients could be enormously useful for hospitals. And ‘risk’ can go beyond just clinical perspective, like this vendor that focuses on malpractice risk. And if its near real-time data-mining based on output from existing systems, even better.

Of course, ideally we would want  (and expect) such intelligence to be inherent in the multi-million dollar enterprise Healthcare IT systems that hospitals buy to record the data in the first place. But most of them are either distracted by industry fads (like RHIOs or Comparative Effectiveness) or bogged down by existing product support to innovate in this direction.

In case you didn’t know, Google.org is the philanthropic arm of Google, and it was formed with the commitment of 1% of Google.com’s profits to address some of world’s most urgent problems (read the famous 2004 IPO letter by Larry and Sergey where they mention it). The site humbly admits that the Flutrends system is experimental. Nevertheless, it’s impressive that in some instances Flutrends was actually predicting flu before CDC.

Of course, not all people who search for flu have flu necessarily, but the power of this analysis comes from the coverage and promptness, not the granular accuracy. The basic idea of harnessing the collective thought (a.k.a. search needs) of the population to predict/monitor health events is fantastic. And this is just the beginning, IMHO. When a population is connected real-time and discussing what they think/want/need, abstracting that information can yield powerful insights- not just for prediction and monitoring, but for most aspects of healthcare (diagnosis, prognosis, news, followup etc).

The concept is applicable to domains outside of healthcare too. Take twitter for example. Twitter is another platform with mass adoption where people are having real-time conversations about what they are thinking/doing. Just look at what intelligent twitter mashups did for getting real-time snow report of the Feb’09 storm in UK or the Dec’08 Ice Storm in New Hampshire. There are health related examples too- the feb’09 salmonella-in-peanut-butter recall could be tracked promptly on a Twitter feed (btw, this slideshare presentation by PF Anderson at the University of Michigan explains ‘Twitter for Health’ in detail. Thanks to Christine Gorman for the link).

Healthmap extracts real-time information from more than 10,000 sites every hour and text mines them for disease and location patterns using bayesian filtering. The interface is clean and intuitive mashup with google maps. Links to the source of alert and a ‘heat index’ (composite score for each incidence based on things like recency of alerts, number of sources etc) are provided.

There are limitations too-  dependance on other sources, unstructured text, lack of integration between sources, not comprehensive, etc. But if you think of it as a free resource that supplements existing public health systems, its a great asset for general public and clinical professionals.

ClearSense aims to help you make sense of your health information by providing the data analysis technology called REDBOX (developed at Bioinformatics Research Center at the Medical College of Wisconsin). I couldn’t find any information around what exactly makes REDBOX unique- it seems like it consists of data analysis models and algorithms that are optimized for health related data (although the devil is in the details for a technology like this). The company behind both of them is Point One Systems, which seems to have spun off from the Research Center.

The sample reports look Web 2.0-ish, with simple interface and layout. The actionable items and alerts are clearly outlined along with tips and educational material. I dont have enough health information in my Google Health account (fortunately I’m in the pink of health) so there was not point in taking ClearSense for a spin on my info.

Although there is minimal info about ClearSense’s unique selling point, the overall concept may actually prove to be useful if/when PHRs take off in future. Hopefully, ClearSense will survive to see that day.

August 2010 Update: Clearsense doesnt exist anymore. Neither does Point One Systems. Oh well.