Health Information Technology Fundamentals

The Journal of AHIMA published an article co-authored by Michael Stearns, MD, CPC in its November 2014 edition.  The link to the index for this edition of the Journal of AHIMA is provided here, however, full access to the article requires AHIMA membership.

The article highlights the differences between the two code sets/terminologies, based on their intended purposes.  Several clinical examples are provided.

SNOMED CT was designed for use in clinical information systems but due to a number of factors it has not been readily adopted in the U.S.  One of the key barriers to its adoption was the requirement that a subset of SNOMED CT must be included (for coding problem list diagnoses) for EHRs to be certified for Stage 2 Meaningful Use.  This has led to a number of stakeholders integrating SNOMED CT into their information systems.  However, SNOMED CT is large and fairly complicated in its structure and design.  It was built using an artificial intelligence construct called Description Logics that allows the system to automatically infer relationships between concepts.  This component of SNOMED CT has been chronically underutilized but has the potential to address a number of challenges in healthcare, including data integrity, segmentation (for privacy), metaanalysis, predictive modeling, clinical research, genomics and proteinomics, clinical decision support, analysis of business operations, and perhaps most importantly clinical research.

The article serves in part as an introduction to SNOMED CT.  For additional information please see the SNOMED CT fundamentals page on this website.   SNOMED CT has a great deal of untapped potential for clinical use, including semantic interoperability.   In particular its ability to define concepts through the use of concept interrelationships.  Of particular interest will be how fully ICD-11 adapts the principles of SNOMED CT and whether or not a migration from SNOMED CT to ICD-11 (or ICD-11-CM) will be seamless or complicated.

For reprint requests of the AHIMA article please contact Dr. Michael Stearns at mcjstearns@gmail.com.

ICD-11 is scheduled for release in 2017, making the delay in the release of ICD-10-CM in 2014 until 2015 potentially problematic.  Further delays in ICD-10-CM, which at this point seem possible, will lead to additional voices recommending that ICD-10-CM be skipped entirely.  This makes some sense from an informatics standpoint as ICD-11 is being modified to make it more applicable to clinically oriented information systems and compatible with SNOMED CT.

If the U.S. goes forward with ICD-10-CM in October of 2015 this may not result a delay in ICD-11 until an ICD-11-CM version is completed.  ICD-11 could be adopted as the terminology for HIT systems in the near future, well ahead of a release of an ICD-10-CM version.  This advantage of this approach would be the following:

1. It would allow for compatibility for disease reporting with the majority of other nations, as ICD-11 will be adopted by the majority of industrialized nations outside of the U.S. within the next 3-4 years.
2. It may allow for improved mapping from a core clinical terminology system (e.g., if SNOMED CT is “updated” to ICD-11) to ICD-10-CM for claims reporting.
3. It will allow advances in the use of terminology to be shared internationally, leading to improvements in data integrity, analytics, predictive modeling, point of care clinical operations, clinical decision making, clinical research, genomic information at the point of care, and other emerging trends.
4. It will allow for improvements in the semantic interoperability between data captured by systems using different languages (e.g., Spanish, English, Mandarin, etc.)

If this approach were to be taken, one might expect to see the rapid development of an ICD-11-CM version in the U.S., unless ICD-11/SNOMED CT can be configured to meet billing and other reporting requirements that are intended features in ICD-10-CM.

The information shared in this article represent the opinions of its author, Michael Stearns, MD, CPC

A recent study (link below) published by the Commonwealth Fund found that smaller primary care practices (small groups were divided into practice sizes of 1-2 physicians and 3-9 physicians) have significantly lower rates of preventable admissions than larger practices (10-19 physicians). This occurred despite the larger practice’s tendency to employ significant greater amounts of continuity of care measures including: the patient-centered medical home, guidelines based reminders, and electronic prescribing.   The study’s findings are limited by it being conducted via a survey of 1,045 primary care practices from July 2007 to March 2009, so its data preceded the wide-spread adoption of EHRs, HIEs, ACOs, and Meaningful Use.

Preventable Admission Rate Comparisons by Practice Size Compared to Practices of 10-19 physicians:

• 1-2 Physicians: 33% reduction
• 3-9 Physicians: 27% reduction

In general, consolidation of practices can lead to higher levels and access to specialty care and supports investments in health information technology, adherence to guidelines, and other benefits to patient care.  However, larger practices also have greater resources that allow them to invest in business practices. This study suggests that one of primary goals of consolidation, i.e., better coordination of care that facilitates keeping patients from being admitted unnecessary,  was not being met between 2007 and 2009.   Further study is needed to assess the impact of practice size and adoption of more modern quality of care policies and initiatives, including EHRs, e-Prescribing, ACOs, PCMH, and HIE.

Author: Michael Stearns, MD

Reference:  Commonwealth Fund Article: Small Primary Care Physician Practices Have Low Rates of Preventable Hospital Admissions

The New York Times article titled “For Big-Data Scientists, ‘Janitor Work’ is Key Hurdle to Insights” is a good read.  It provides an excellent overview of the challenges associated with making sense out of massive amounts of data being generated by disparate information systems.  The dramatic increase in the use of EHRs and related technologies in healthcare have fueled a data explosion that is unprecedented.  An avalanche of information has been coming from thousands of HIT systems using different languages.  Despite efforts going back over 30 years, the vast majority of data that is attached to any type of code is claims data (e.g., ICD-9-CM codes).   This type of structured data is notoriously unreliable in clinical care and only represents a small fraction of the data needed to make a difference in healthcare research and population management.

SNOMED CT, LOINC and other reference terminologies/ontologies attempt to give structure and meaning to information at the level of the concept.  The challenges faced by big data scientists would be markedly reduced if a greater amount of healthcare data was stored as codes using these types of terminologies.  Progress has been slow but Stage 2 Meaningful Use does offer hope that the adoption of SNOMED CT will increase, and that EHRs and other HIT systems will begin to adopt codified terminologies as the core language of their information systems.  These efforts have the potential to transform healthcare and genomic research.

The article can be found here: “For Big-Data Scientists, ‘Janitor Work’ Is Key Hurdle to Insights”

Author of this post: Michael Stearns, MD