Anticlue

Usability and Patient Care

Over the years, the health care industry has been increasingly concerned about privacy and personal health information. The other item is one of the most critical activities during system processing is the decision to choose the right patient. The ability to intelligently select the correct patient based upon data in the screen is critical to computing systems enhancing and assisting in the patients care.

That said, I read Jacob’s entry on usability and the practice EMR yesterday. The comparison of the searching functionality of google compared to an EMR, was concerning. This is what I have understood from the entry. There is a time-loss for the family doc to enter the patient’s name. Jacob thinks having the google suggest search mechanism helps with usability. This inconvenience needs to be fixed, but at what cost?

My interpretations are:
First, over the past several years, there has been regulations passed on the privacy of health information. Those security and privacy regulations should be a primary concern of any application in healthcare, but especially an electronic medical record. Dumping the Master Patient Index, MPI, into the browser to have a javascript typing search is just looking to not comply.

Second, over the past several years performance of a system is critical. Again dumping the MPI into the browser to have the javascript typing search doesn’t help out a lot with performance. Also having a db search done at every key stroke doesn’t help with performance. I don’t know what would be more frustrating, waiting for the response between keystrokes, the wrong name popping up, or typing the name out. My guess would not be typing the name out. By the way, I’m 99% certain neither mechanism was utilized for the google suggest.

The next item is that names are not always spelled correctly. For instance when saying my last name, Nielsen, I’ve seen it spelled Nelsen, Nelson, Neilsen, Neilson, etc. I would really prefer not to be treated based on the information in Alysse Neilson’s EMR. Also transposing of numbers is exceptionately easy and frequently done. This is why the name and mrno is not enough information to uniquely choose a person. I previously posted about the information needed for an MPI here. This information assists in the decision point of choosing the correct patient.

So what would need to be done for this issue. First, I recommend Jacob put together a list of issues for his practice EMR implementation. Then I suggest Jacob offers the list to his IT or vendor group. Give them an opportunity to hear him, and work with him. (Notice I said with) Rank each issue with its necessity for live, include information on the level of effort needed to resolve, the criticality of the issue and the risk of the issue. Work with the vendor group to resolve the issue based on the needed timeframe for live.

There may already be some solutions, perhaps typing a couple of letters and having a soundex search return the subset with MPI identifiers. Autocompleting is fine for googling, but what about when someone’s health is at stake?

Also if this issue was that critical to the usability of the application, why wasn’t that a part of an RFP functional requirements?