Health IT Quality

Surescripts has published a newer version of the previously published Quality Guidelines document that outlines the expectations and requirements of a high quality electronic prescription.  This new version provides additional details on some of the key areas of clinical focus and concern observed during our ongoing review of new electronic prescriptions. This newer version of the Quality Guidelines attempts to set the measurement standard against which electronic prescriptions can be assessed for accuracy and completeness.

I had a chance to read the recent NIST publication on EHR usability.  If you have not seen this document, I strongly suggest you read it.  There are many concepts that are parallel to the work we have done with e-prescribing quality.  In particular, I like the terminology and organization they have.  No need for Surescripts to put together something like this on our own.  As much as possible, we will try to refer to this document when we talk about the impact that e-prescribing user interfaces have on the outcome, which is a safe, accurate, unambiguous prescription medication order.

At Surescripts, we have a measurement system of the output of the e-prescribing user interfaces.  By looking at issues and areas for improvement, and tying that back to the source, we think we can develop a cause and effect model for the transferral of digital health information.  Specifically, we have seen good user interfaces which derive good results.  We have also seen weak interfaces which result in poorer outcomes.  I don't think users intend to create problems.  Seeing a connection between cause and effect is critical in improving safety and outcomes.

I applaud the work of NIST on this publication.  I look forward to referencing it and working with the information in this document.

Once you've had a chance to look at it, please make a comment below and let's see if we can all find a way to incorporate this work into what we do day in and day out.

Not every problem that we see with an electronic prescription is a black and white safety issue.  The vast majority of the issues we see are inconsistencies or inaccuracies that do not meet the Guidelines, but are not absolutely going to cause a patient safety issue.  The point is that the inconsistencies increase the possibility that an error will result, because it requires interpretation, analysis, study, or re-keying of data.

There are two outcomes from this observation.  First, getting attention on problems that are not themselves safety issues is challenging.  Everyone wants to fix patient safety problems.  I have a harder time getting attention on problems that are absolutely quality problems but not outright safety issues.  Second, this can account for the different opinions on how many errors there are.  Depending on how you define "error", two different people can look at the same prescription and one will call it an error while the other will not.

Is it really quality versus safety, or is it safety in the context of quality?

Do you have an opinion on this subject?  If so, please add your comments to this discussion.

I've been asked to explain how we produce the clinical quality scorecards that we send out.  The process is fairly simple.  We start with a selection of electronic prescriptions.  Usually we will just select a time range, like a day or a month.  We never look at PHI.  In fact, PHI is not available in our engine when we pull data.

The data is then run through the Surescripts Quality Engine.  I'll leave the details of that process to another posting.  Suffice it to say that the engine does its best to try and classify the prescriptions into those that are perfectly fine, and those that have suspect content.

The data is then reviewed by a Certified Pharmacy Technician, who validates the decisions that the engine made.  The technicians then review the content that the engine cannot reliably score automatically.

We measure the content against the Guidelines, and score them using the Terminology we have developed.

We then aggreagate the findings into a scorecard.  Depending on how we are going to use the output, we prepare spreadsheets, powerpoints, emails and custom reports.  Sometimes the data is broken down by vendor, or by practice doctor or pharmacy.

A recent article published in the Journal of the American Medical Informatics Association (JAMIA) studied the occurrence of errors in prescriptions that were delivered to retail pharmacy.

Once we saw how the study, the press release and the resulting media coverage equated the findings to e-prescribing, we contacted the authors.  We were able to confirm that the study did not look at any electronic prescriptions. The study looked only at computer generated faxes and printouts. Prescriptions that are faxed or printed are not subject to all of the certification checks and qualitative requirements associated with e-prescribing. The authors agreed that the results of the study should not be generalized to e-prescribing or linked to errors in e-prescriptions as there are clear and significant differences between these prescription transmission methods and the terms and technologies are not interchangeable.

To clarify these points, we worked with the authors on a joint statement that was released on Friday July 8, 2011.

The study also underscores how there are numerous methods of measuring and defining prescription errors. In our statement, we issued a call to harmonize these approaches so that researchers and health care professionals can compare results between studies – more effectively and without ambiguity. Doing so will create a more consistent and reliable record of measuring improvements in prescribing accuracy, efficiency and safety.

Please comment below on the study and/or the joint statement.