Data Challenges on the EHR Agenda

By John Glaser

While preparing to adopt or expand electronic records, providers shouldn't overlook the quality of the data these systems will contain.


John Glaser

The health care information technology portions of the American Recovery and Reinvestment Act (ARRA) have led to heightened interest in the adoption and effective use of electronic health records. Given the importance of improving many facets of care and the magnitude of the stimulus funds, this attention is appropriate.

Most of the discussion about advancing EHRs centers on areas such as meaningful use, certification, interoperability and regional extension centers. Underneath these topics is the industry's focus on the software application called the electronic health record. For example, when the industry talks about adoption and effective use, it is referring to the EHR software application. And when the industry discusses interoperability, it is focusing on EHR applications being interoperable with each other.

This focus on the EHR software application should not distract us from also concentrating on the data in the EHR. Both the near and intermediate terms of the national EHR agenda pose several data challenges:

Large-scale information exchange among health care entities raises data management questions for both the senders and recipients of data. For example, under which conditions can data from one organization be used for clinical research by another organization? And if one organization needs to amend data it has exchanged with others, how is that amendment propagated?

Changes in privacy regulations will require improvements in data access controls and tracking data movement within and between organizations.

Broad EHR adoption will open the door to a diverse set of secondary uses of data for clinical research, care improvement, population health and post-market medication surveillance. Early efforts to leverage EHR-based data to accelerate clinical research, and to dramatically improve the efficiency of post-market surveillance, show promise but have also exposed data quality problems.

Widespread EHR adoption may accelerate use of personal health records. Patients using PHRs will likely increase their contributions of data, such as measurements from home monitoring equipment, to their EHRs.

Federal efforts to improve safety and quality will increase the use of clinical decision support. This, in turn, will heighten the focus on data that drives support algorithms and rules. For example, if the problem and medication lists are incomplete, the potency of drug-drug interaction warnings in the EHR can be seriously diluted.

This is not an exhaustive list of the issues that we will confront in the years ahead. Nor is ARRA the only reason to focus on data; ICD-10 implementation and private payer pay-for-performance initiatives are also factors.

New Considerations

The industry's understanding of these looming issues and challenges—and how best to address them—is imperfect. Despite this, the following ramifications are clear.

The breadth and depth of provider data quality management efforts must increase. In general, most providers focus on the quality of a very narrow subset of data required for reimbursement, targeted care improvement initiatives and external reporting.

In the years ahead, quality management efforts will need to encompass a broader range of data, including data from patients and from other organizations via a health information exchange. Providers will also need to consider applying quality management processes to more unusual forms of data, such as the knowledge assets of EHR templates, computerized provider order entry order sets and clinical decision support rules.

In addition to expanding the breadth of data management activities, providers' data quality management activities will need to be more rigorous. Today's approaches to ensure that the meaning of a particular data element is unambiguous, to document consistency of source application systems and to monitor data quality may no longer be adequate. The degree of data management prowess that is adequate for reimbursement may not be adequate for effective use of clinical decision support. Problems with clinical data quality will severely hinder the value of a health information exchange; receiving providers may not use data from other organizations if they don't trust its comprehensiveness.

Provider organizations have to amend data use and management policies and procedures. Virtually all providers have developed policies and procedures that govern internal and external use of patient data. These policies and procedures cover territories as diverse as approving new clinical forms, defining time frames for completing documentation and determining conditions for release of information.

The policies and procedures will need to be adjusted to reflect changes in privacy law, the specifics of meaningful use and the growth in commingled data that will come from health information exchange use. For example, if a provider's clinical data repository contains data that originated from dozens of other providers, policies governing secondary uses, such as improving care operations or conducting clinical research involving anonymized patient data for genome association studies, will be needed.

The definition of the medical record will become more fluid. To the degree that the medical record is defined as data that a provider accessed or had access to during the course of providing care, its boundaries will expand. The definition will grow to include data received from other providers, patient-contributed data and computer-generated clinical decision support warnings, alerts and reminders.

This broader definition has malpractice and care implications. The industry will need to reach consensus on a definition of the medical record that is no longer rooted in a paper chart and confined by the provider's organization. In many ways, this new definition is long overdue.

The Years Ahead

To improve care through the broad adoption and effective use of EHRs, we must focus on the EHR application software. Applications impact workflow, possess features that support documentation and e-prescribing, and must be installed on hardware and software platforms.

We must also have a parallel data focus. Data will be exchanged and used to assess care performance. Privacy protections are centered on data.

The years ahead will require that providers review and significantly change their internal data management functions to encompass broader scope, increased rigor, and a new definition of the electronic health record. Ensuring that we address data centricity well will require a collective, multistakeholder discussion.

John Glaser is vice president and CIO of Partners HealthCare in Boston and a regular contributor to H&HN Weekly.

This article 1st appeared on June 8, 2009 in HHN Magazine online site.

To respond to this article, please click here.