Blogs

Like distributed capture, and SharePoint; document imaging in the healthcare / insurance space is one of those trends that is always at the top of the interest list, but not very mature.  The demand to find solutions for reducing paper in healthcare is driven by the need for efficiency, government rebates, and keeping a head.  However, companies have struggled to determine the best technology and the best use of the best technology for years.  The use cases I’m explorer here are:

Medical billing:  The automation of paper medical billing documents for both insurance and practices.  The document types in this area are HCFA/CMS1500, UB92, ADA forms on the insurance side, and EOB on the practice side.  Most often a major component of such a solution is the marrying of paper forms with EDI (Electronic Data Interchange) equivalents.

Electronic medical records ( EMR ):  Increasing the access to medical records documents both historic and current.  The goal is to give customers better access to their information, and increasing the efficiency of practices.  The types of documents in this group are very widespread.  Because of this the demand is not so much on data extraction, but data classification.

In both use cases the adoption short comings have been:

Regulation:  HIPPA compliance and the like are a benefit to the sustainability of these technologies, but not to their adoption.  Regulation from my perspective has the biggest impact during the evaluation of the technology.  Regulation forces organizations to consider canned demos versus demos on production documents.  This is risky, and slows down the process of choosing a technology.

Hope that paper will die:  There is a large group of organizations hedging their bets that the documents will turn to EDI.  EDI removes the paper and makes automation much easier.  While this will happen slowly, and is more an issue of standards than anything, there will always be the handful of documents that don’t get converted.  Because of the shift, these paper documents are going to be the most costly.

Underestimation of complexity: EOB’s are the most difficult paper document to automate; the other documents in healthcare are not far behind.  IT departments approach the automation with great enthusiasm.  Prior to deployment estimations of effort is given, which are not backed by reality.  When solution exploration begins, very often the budget and effort required is startling to project owners.

Incorrect focus:  The focus for most insurance companies and hospitals especially has been on the resulting file format HL7, 835 etc. from automation.  However, the reality is these file formats like all file formats are trivial compared to the effort it takes to get the data to generate them.  This focus has caused organizations to pick solutions not on their quality of capture, but on their ability to export a file format.

The best ways to address these short comings are:

Start off small: Find smaller projects that are very desecrate and require less technology.  For example, in an EMR scenario instead of setting a goal to image all medical records, file them correctly, and extract pertinent information to associate with a patient record.  Focus on just the scanning and classification of the documents.  This augmented by a quality viewer to get the data will provide value with a greater chance at success.  When successful, move on to the automated extraction of data.

Focus on what will automate easily:  Pick the documents that will automate easily.  This will be 300 DPI drop-out HCFA/CMS1500 forms, or specific insurance EOBs that are clean and generation one ( meaning not a copy).  The technology is advanced enough to take countless variations and make sense of it, does not mean you should try it.  Hone your focus to the documents that you know you can automate with great accuracy, and do not require tremendous effort.  Your first deployment’s set up should be minimal ( 2 to 3 weeks ).  After you have an understanding of success, and a measurable return, scale to more complex documents.  What kills organizations is trying to automate the complex documents upfront, and later as they scale forgetting about regression testing.

Research: Investigate potential solutions prior to establishing a project or goal.  Understand what is possible in what scope. 

Be open to change:  Be open to first not automating everything.  Only automate what will give you value. And second making changes in how you do things.  This could be scan settings, document prep, and even working on the source of the documents to increase quality.

The technology for EMR and medical billing paper automation is clearly there.  The complexity is substantial when looked at the whole picture, but when taken in pieces the path is pretty clear.  For medical billing, EDI is coming at a rapid pace, and soon will be a large percentage of the transaction volume.  For EMR, paper will always be a challenge, and the highest priority should be put on classification and quality, not data extraction.