Development of Risk-Adjusted Outcome Measures in the EHR Environment

Historically, heath care quality measures have focused on processes, not outcomes. Outcome measures quantify the result of an individual’s encounter with the health care system and thus are thought to be better measures of clinician performance and quality of care efforts. Challenges with implementing health outcomes measures include difficulty obtaining the data from electronic health records (EHRs), extended length of time between a process change and measurable impact on health, and the need for risk adjustment. Risk adjustment is the process of taking patient-associated characteristics—such as co-morbidities, severity of illness, and physiological status—into account to enable fair comparisons of outcomes across different patients, treatments, providers, or populations.

This project held an invitation-only conference to examine methodological issues related to the identification and development of outcome measures and their accompanying risk-adjustment models in the EHR environment. The scope of the conference was to:

• Identify ways that the EHR can be used to develop new risk-adjusted outcome measures across health care settings. 
• Identify ways that the EHR can be used to enhance current risk models that have been developed for existing outcomes measures. 
• Use the results of the conference to form the basis of a demonstration project during which risk adjusted outcome measures will be developed and tested using EHR data. 
• Widely disseminate the knowledge and information gathered from the conference through a white paper. 

The conference was held March 14, 2012. Sixteen experts in outcome measure development, risk adjustment, and health information technology participated, along with one consumer stakeholder. Sessions were presented on outcome measures; EHR technology and standards; EHR implementation; and risk adjustment. Facilitated brainstorming sessions were conducted to address the use of EHR data for risk-adjusted outcome measures.

Although meeting participants discovered they are further away from identifying ways that the EHR can be used to develop risk adjusted outcome measures than anticipated, they were optimistic that data derived from an EHR for outcome measurement purposes will be feasible in the future. Currently, however, there are significant barriers to be overcome in order to accomplish this. These include lack of eMeasures definitions, difficulty in data capture, lack of standards, and lack of robust interoperability. Attendees suggested a number of demonstration project ideas that could form the basis for future study in this area. A white paper on conference discussions is being written and will be publically released in the future.