Standards for “Meaningful Use” Before Semantics

March 23, 2011 by
Filed under: Business Rules, Data Governance, Recommendations 

I had done some talks on health care business intelligence and data quality. This morning I was pointed to a short article about health care and data in which the claim was made that the “health care field is fertile ground for semantic tech.” In the article, a reference to NIEM (the National Information Exchange Model) was made suggesting a “greater use of semantics.” Former Fed and now CTO at Accelerated Information Management Michael Daconta is quoted as saying that “the federal government’s ‘meaningful use’ directive, which focuses on the adoption of electronic health records, calls for decision support.”

“Meaningful use” is the term used to refer to the push for high quality data managed as electronic health records. According to the CMS web site,

The American Recovery and Reinvestment Act of 2009 specifies three main components of Meaningful Use:
1. The use of a certified EHR in a meaningful manner, such as e-prescribing.
2. The use of certified EHR technology for electronic exchange of health information to improve quality of health care.
3. The use of certified EHR technology to submit clinical quality and other measures.
Simply put, “meaningful use” means providers need to show they’re using certified EHR technology in ways that can be measured significantly in quality and in quantity.

I don’t want to step out of line here, but having read through a number of documents about the criteria for meaningful use, I think there is a long way to go before we get to decision support and semantics. Basically every practitioner and every health management organization is being incented to produce health records in structured electronic form. Demonstrating meaningful use means observing quality criteria for 25 meaningful use objectives. Some are core objectives and you might say the others are electives – for hospitals, et. al. you have to meet 14 core objectives and can choose 5 from a list of 10 additional ones.

The simplest might be measure 6, recording demographics. For this one, you need to record the following as structured data:

(A) Preferred language

(B) Gender

(C) Race

(D) Ethnicity

(E) Date of birth

(F) Date and preliminary cause of death in the event of mortality in the eligible hospital or CAH

If more than 50 percent of the records associated with the admitted patients have all these demographics records as structured data, the criterion is met. However, “structured data” is a pretty broad concept, and there are very few directives indicating standards applied for any of these demographic attributes (race and ethnicity is, but nothing for the others).

So here is the scenario: Hospital A records patient demographics using their own data model and data elements. Hospital B records patient demographics using their own data model and data elements. Both share their data with CMS. Neither use the same data formats. Before CMS can even do any decision support or analysis or even reporting, they have got to absorb the data. Where are the standards for exchange?

Yes, there will be an opportunity for semantic technology. But I think that it’s going to be needed at the point of data sharing or intake at CMS, which is way before the data can ever be used for analysis.  Hopefully someone at HHS is thinking about standards for data elements used for meaningful use.


2 Comments on Standards for “Meaningful Use” Before Semantics

    […] Loshin, president of Knowledge Integrity, shares some insight about meaningful use, incentives, data integration, and the need for standardizing data element structures for […]

  1. James P on Tue, 3rd May 2011 4:57 PM
  2. No David, it is going to be needed at the point of capture. This data is used in operational decision making long before it is used in BI. If the data does not support inter-operability between hospital A and B, it is no good. And if each hospital has 10 different applications and many versions of the same data, this is unacceptable; but today’s reality. This is a dangerous, high risk proposition, and part of the mess of healthcare.
    Healthcare has a legacy of no enterprise data architecture – at any level of enterprise. The EHR is a myth because while anybody can create one (and they will) there are no rigorous and enforced standards. Healthcare doesn’t understand data; that is they don’t make the connection between data quality and healthcare decisions and operations. They think they just have to create records and keep records. Software vendors don’t get and don’t care to be too concerned about data because messy data creates more work.
    Payers care about some aspects of the data because they want to adjudicate and re-price. But every payer has a different set of requirements. Again, unacceptable. The provider data has to be edited and translated at a switch for every different payer’s requirements.
    We have 3 different EDI standards for healthcare claims and there is only need for one. The HL7 work was focused way too low for meaningful standards – it is based on data modeling.
    Semantics are critical, and not just for BI, but for patient care. Data standards and Data quality start at the point of capture and DQ services need to be standardized along with the data structures and the data content. If it was my EHR money, that is where I would put it. (oh, wait it is our money)

Tell me what you're thinking...
and oh, if you want a pic to show with your comment, go get a gravatar!