Meaningful Use lab results: laboratory integration
Ever since the ONC made its recommendations to CMS for Meaningful Use and Certification of Electronic Health Records EHRs, we have been commenting on several aspects of the 25 criteria categories needed to be eligible for HITECH incentivesstarting in 2011. We have commented on the question of Certification and stand-alone billing systems, and on challenges around interoperability, particularly with respect to statewide and regional Immunization Registries.
An additional area of challenge for establishing an eventual “plug-and-play” national health information system pertains to laboratory data. Incorporating clinical lab-test results into EHRs as structured data is one of the Meaningful Use criteria (criterion 10) – physicians must be able to demonstrate that 50% of clinical lab tests that were ordered (whose results were numeric or +/-) are incorporated as structured data.
In order to achieve this, a common “language” must be implemented by all laboratories, so that EHRs can import these results in a systematic fashion. The specific vocabulary specified in the Certification documents is to use a coding system called LOINC. There is a standardized LOINC code for each different lab test type, specimen type and specific methodology – there are about 40,000 different LOINC codes currently defined for lab tests.
In theory, this is great. However, two issues have emerged when trying to implement standardization around LOINC codes.
1. Not everyone uses LOINC codes. Major commercial labs (like LabCorp and BioReference) use LOINC systematically; Quest uses LOINC regionally but there are regions within Quest that have not yet been moved to LOINC (it is on their roadmap to achieve this in 2010). But that only accounts for a portion of all labs ordered by physicians. Local hospital labs are very inconsistent in their adoption of the LOINC system, and often use their own internally-derived coding to define lab tests they do. In fact, some legacy Laboratory IS systems widely utilized do not even have a field placeholder in their database for a LOINC code.
The LOINC organization makes available a tool (called RELMA) which local hospitals and other labs can use to convert their internal coding methods to the LOINC standard – however, this implies that local labs will have the desire and ability to do this. The result is that a significant proportion of lab results reported locally around the country are not currently on the LOINC standard.
2. The LOINC system is exceedingly granular. For the purposes of aggregating and trending data across multiple sources, data needs to be grouped into “buckets” of like-types. For example, if trending blood glucose for a diabetic patient, one will want to pool together all the results for blood glucose for reporting. There are 68 different LOINC codes for glucose, many of which are seemingly the same test.
The LOINC system is a bit reminiscent of the medication e-prescribing domain, in that the basic nomenclature for a drug to be prescribed is its NDC code (which specifies medication, dosage, delivery type, and manufacturer), which is similarly very granular. However, for medications, there are existing companies that provide databases that aggregate NDC codes into logical groupings (which is how physicians think of prescriptions) – by medication, dosage and delivery type (e.g. tablets, suspension, etc.). In the realm of lab tests, however, there are no such existing databases that aggregate LOINC codes into logical groupings – this is a growth niche opportunity for companies that would want to take on that task.
Trending of lab values is important for patient care. It is also the basis for reporting of a number of PQRI and HEDIS quality metrics, and is the basis for other Meaningful Use criteria (criteria 12 and 14). The challenges of (1) collecting lab data in a standardized fashion using the LOINC coding system, and (2) aggregating those values into logical “buckets” that make sense for quality metrics and reporting, are surprisingly difficult to address – even with something as seemingly-straightforward as lab test results (nice, structured, numeric things).
Traditional, legacy, locally-installed EHRs may have an easier time accomplishing this kind of integration if it is done with one or two local labs – the integrations would need to be set up point-to-point, and would likely involve IT consultants to achieve, but then the local connection is made. However, for emerging web-based EHRs, such as Practice Fusion, with a national scope, point-to-point integration is impractical. In order for the laudable vision set forth by the ONC to establish a truly interoperable nationwide health care delivery system, all parties – including all the local and hospital labs – need to “speak the same language” and, as far as laboratory testing is concerned, adopt the same system. All of us – EHRs and laboratories – need to make the needed changes to accomplish Meaningful Use.
Robert Rowley, MD
Chief Medical Officer, Practice Fusion, Inc.