The Future of Documents

 The Stimulus package is definitely a great incentive for getting small practices and large hospitals to move towards Electronic Medical Records – despite a pretty high upfront cost of around $44000 per physician to install a new electronic health-record system. Daily Finance has an interesting article and interview on a new trend: ad-sponsored online health records.

Practice Fusion, a small start-up, has an interesting approach of making that service free but ad-sponsored. Their software is web- and cloud-based (partners of Salesforce.com), meaning doctors don’t have to worry about setting up the software. Even better, it can be free, provided doctors agree to have ads appear on their record system.  Practice Fusion provides interesting capabilities, like automatic charting, patient management, ePrescription, scheduling and billing.

One thing that leaves me a bit uncomfortable with many of these proposals is the gap between past and present (paper) and future (full digital), and the deliberate avoidance of the hardest problem - getting legacy paper records accessible in the new system. Sure, new paper documents can be scanned and imported as images, but what about the legacy volume of documents still sitting in folders? How can you extract and inject them into an electronic Medical Record system, while making sure this information can be searched, accessed and retrieved easily?

I visited one of our customers recently, who has a huge warehouse of over a million Medical Records folders -  scary experience, especially when thinking that my life might depend, one day, on the speedy access to the right information contained in that 200 pages folder, sitting with another million folders …

So how do you intelligently scan the legacy medical record and recreate an intelligent, electronic version is navigable, searchable, and brings as much information to the doctor -and hopefully more- as the physical paper record? That is, to me, the toughest problem. I’ll be touching on some of these aspects in the future.

Most major companies are suffering from Information Overload, and in particular dealing with very large amounts of incoming paper and digital information. This information needs to be manually processed, before it can be delivered to the right knowledge worker in your organization – and in many cases, it can take up to 15 days for that information to be delivered to the right person. So think about your typical customer request – by the time you get back to her, she has moved away and is presumably very dissatisfied with your company. Turnaround is key.

This is why Xerox is working on technologies to address Information Overload. One such technique was mentioned in this article in The Times magazine ”Confronting the information overload”. Called the Hybrid Categorizer, this technology automatically sorts and classifies documents. As opposed to existing techniques, which solely rely on “visual” (e.g. shapes) or “textual” (words) to recognize a doctype, Hybrid Categorizer takes into account both the visual and textual information. 

Plus it fully leverages Machine Learning – meaning it “learns” what characterizes each doctype, as opposed to requiring a human to “teach” (often) subjective rules. It is therefore capable of achieving a quantum leap in the Automatic Document Recognition it can achieve – this with minimal setup and errors.

I’ll cover some real use case studies of Hybrid in the future – this technology is used in many applications including the Digital Mailroom, which is part of Mail and Distribution Services. For more information click on this brief illustrative video:

EMC, the leader in Entreprise Capture technologies, recently announced the next step towards a more efficient Document Processing infrastructure – their new InputAccel and Dispatcher version 6 establish a new benchmark for Document Processing and Capture.

The most interesting part of the announcement is probably their adoption of a standard Service Oriented Architecture (SOA). This will enable much simpler integration into entreprise applications, allowing individual scanners and multifunction to become portals at different steps into your business processes – not only as the entrypoint. The power of Dispatcher technologies becomes available to any office device, not only to high-end production scanning centres. Plus best-of-breed 3rd party technologies can also be integrated seamlessly – no more VB-scripting and direct dll invocation.

The next logical step might be to host that infrastructure on “the cloud” to make Document Capture 2.0 not only available to each office device in large entreprises, but also to SMBs. Sounds promising!