As the Chief Technology Officer of a large healthcare operation responsible for retrieving records from over 33,000 locations per month, I have a deep interest in making sure that our customers receive precisely the records that they have asked for.
While this sounds easy, in practice it is very difficult, because even when we make extremely specific requests of providers there is a tendency for them to return additional records that are not relevant. While we don’t know all the reasons for this it’s probably fair to assume that the front office staff who are sending the record are probably not sufficiently trained to make the necessary clinical judgments to properly curate the patient’s information, and as such they feel more secure sending the entire patient file.
I’ve heard some people suggest that the right way to deal with this problem is to wait for a fully electronic system and then have the doctor or nursing staff make the selection of the records they are interested in. I’m really concerned about this approach because it sounds like the tragedy of the EMR all over again. If your average patient must be seen in 20 minutes do we really want our medical professionals, both nurses and doctors, to spend yet more time on the computer looking for records? Or will we be outsourcing this operation to companies like mine even in a fully digital world?
Let me give you a concrete example of how we handle these problems today.
Suppose that we have a patient named Jane Doe who is unfortunate enough to have breast cancer, and has been sent for a consultation at Acme healthcare. The intake coordinators at Acme have worked with eHealth Technologies to build a set of clinical templates for the various types of clinical notes such as labs, progress notes, history and physical, radiology, pathology, and so forth. They decide to enter a new request for Jane’s records.
Our service center staff use the information provided by the intake coordinators to go out to a set of medical centers, and make specific requests such as “give me all Jane’s CT exams for the past five years”, “give me all Jane’s pathology for the past six months”, or “give me all radiology reports relating to Jane’s ultrasonography of the right breast”.
While these requests are specific and clear, the information that is returned rarely meets the specification. We may see plain films of the chest that were used for pneumonia screening during a round of chemotherapy, or history and physicals for a visit that was related to a urinary tract infection, and so forth. If we left all of this material in then our customer would be far from happy, so our staff are trained to make judgement about how to separate records into what we call the “requested” and “non-requested” categories.
In some cases this is an easy process, but there are certainly clinical judgement calls that must be made and it is important to have a constant line of communication with the customer to ensure that we are correctly categorizing the records. In the example above an ultrasound guided biopsy of the breast is not the same as a radiology read, but here we must always err on the side of caution.
It is important to understand that most records are sent as a concatenated mass, and our service center must do significant work to break all records into discrete encounters and determine their data service and record type. Once these elements are known we must categorize the records by note type, before our service center starts the process of determining what was requested, and what was not requested before delivering both sets of information to the customer as separate documents.
This process of filtering clinical information is difficult, time-consuming, and requires significant training, skill, and technology to ensure that our customers get the right information right when they need it.