Lost in Translation? Clinical Decision Making and the Need for Lab Data Standards

By Gai Elhanan, MD, MA
02:35 PM

The HITECH initiative and the promise of effectively coordinated care are fundamentally based on the adoption of standards as an integral part of the larger adoption of healthcare information technology. Numerous types of standards are being promoted, including messaging standards, secure communication standards and data standards. But perhaps, some of the most important standards are those that are not being enforced.

As part of the daily discussions regarding ways to ensure the success of HIT initiatives, the talk frequently turns to problem lists, medication lists, allergies and similar topics as being the cornerstones for delivering effective, coordinated care. Clearly, the quickest route to successfully demonstrate cost savings is to reduce preventable readmissions and emergency department visits. Accountable Care Organizations have jumped on the bandwagon with recommendations around care coordination, as well as discharge and medication reconciliation. And, while there is no denying that these areas are extremely important, missing from these discussions is a key component of the healthcare mix – clinical laboratory data. Somehow, that part of the healthcare equation has been taken for granted.

Clinical laboratory data comprises the bulk of an individual’s health record, with estimates ranging from 60 percent to more than 90 percent of the data – regardless of whether the health record is electronic or paper-based. Lab data is clearly essential to the medical decision making process and is estimated to be involved in 60 to 70 percent of all medical decisions. Yet, standardizing this data has not been addressed with the same intensity as other elements of the health record.

Making lab data meaningful

When most people think of ambulatory clinical lab data, they think of corporations like Quest Diagnostics and the Laboratory Corporation of America (LabCorp). Yet, how many people know that Quest and LabCorp, together, are responsible for only about 21 percent of the clinical lab data generated today? And this number is slowly shrinking. In fact, the vast majority of clinical lab data in the U.S. comes from a multitude of hospital labs (~8,000) and independent labs (~6,000). And while large, national labs such as Quest, LabCorp and a few others are doing an admirable job of providing standardized results to their clients, the vast majority of lab results data is being generated by outdated laboratory information systems, utilizing proprietary test codes.

So, how does all of this fit into the ongoing quest for data interoperability? Results generation is, in reality, only the tail-end of the problem. Many of the available test ordering systems are also less than optimal. When considering EHR certification and Meaningful Use challenges, it’s easy to assume that certified physician order entry (CPOE) will solve this challenge. However, Stage 1 Meaningful Use guidelines address order entry primarily in the context of medication orders. In addition, many electronic lab orders are eventually printed and manually re-entered into laboratory information systems, do not undergo the necessary checks such as Medical Necessity, and are coded by CPT codes – all of which not only reduce productivity and increase error potential but also limits the accuracy of the data.

To LOINC or not to LOINC?

CPT (Current Procedural Terminology), in fact, is a misnomer. It is neither a terminology, nor is it truly current when it comes to clinical labs. CPT codes were designed for billing/payment purposes and not for clinical use – not to mention that they are a proprietary coding scheme. Currently, 1389 CPT codes exist for clinical labs, although that numbers seem to be shrinking with recent versions. In comparison, the Laboratory Observations Identifiers Names and Codes (LOINC) system includes more than 44,500 codes for laboratory data, and is growing. Despite this trend, in the final rule, CMS says: “We clarify that we do not expect Certified EHR Technology to natively (or internally) support LOINC in its entirety, which is why we do not believe that it is necessary to specify a subset of common LOINC codes” and “We do not expect, … , that Certified EHR Technology will have to crosswalk or map internal or local codes to LOINC codes.” If we can’t expect EHRs to support a LOINC standard, why should we expect labs to adopt the standard?

Many of the proposed quality and outcome measures, PQRIs and data analytics that are expected to be implemented to improve population health are dependent on quality data. CPT codes, or proprietary codes, are not sufficient. And, while voluntary adoption of LOINC by vendors is a noble cause (and a necessary one, in my opinion), this is not likely to happen at a significant rate unless the consumers of healthcare technology put pressure on vendors. Healthcare providers and organizations that purchase and implement any HIT technology should not only verify that the technology is “certified,” but also that it is fully LOINC capable in receiving, displaying and transmitting clinical lab-related data. Because, at the end of the day, clinical data helps make informed healthcare decisions possible.


Gai Elhanan, MD, MA, is chief medical information officer at Halfpenny Technologies, based in Blue Bell, Pennsylvania, and is a research professor in the department of computer science at the New Jersey Institute of Technology in Newark, New Jersey.