Improving the Quality of Quality Measurement
Mass. Veterans Epidemiology Research and Information Center (MAVERIC)
VA Boston Healthcare System
Leonard.Davolio@va.gov or firstname.lastname@example.org
It seems overly obvious that a prerequisite of improving healthcare is the ability to measure the care being delivered. Unfortunately, most electronic medical record systems were not designed to support quality measurement. As a result, attempts to measure quality must contend with disparate and often unstructured data of questionable quality for care measurement. These challenges lead to few widely implemented quality measurement programs, many of which are limited to the number of charts that can be manually reviewed or to the relatively few measures that can be informed by existing structured data.
From an informatics standpoint, automated quality measurement presents a great opportunity to explore a number of interesting challenges while contributing to an important area of healthcare improvement. For example, a real time or near real time quality measurement system must contend with access to EMR data, data quality issues, unstructured data, analysis, and the presentation of data. We've attempted to address these challenges in two important areas; primary care and colorectal cancer (CRC) surgical care.
For our project in primary care quality measurement we worked with Imran Faizi, a leader in improving the quality of veteran care at the VA Hospital in Chicago. There we automated the collection of important HEDIS-defined measures by tapping into the VA's clinical reminder system to provide role based reports to primary care providers and aggregated views to department directors. For colorectal cancer surgical care we used the VA's Medical Domain Web Services (MDWS) to access pathology reports, machine learning (using ARC) to identify CRC post-op pathology reports, and natural language processing components (UIMA-based modules) to extract the number of lymph nodes examined and the number positive for cancer. The VA has an enterprise license for Microsoft products and Sharepoint is FIPS compliant and VA-approved so we used it as a platform to present dynamic, role-based reports and email notification through Active Directory.
We've created two videos that demonstrate how we implement end-to-end quality measurement. [Video 1] shows how data is retrieved from the EMR (in this case, VistA) and how it is prepared to inform quality measurement. [Video 2] explains how the prepared quality measurement data is delivered to various stakeholders.
Preparing the EMR data [Video 1]
Presenting the Quality Measures [Video 2]
This work is made possible by funding from the Veterans Administration Cooperative Studies Program (CSP) and Health Services Research & Development through the Consortium for Healthcare Informatics Research (CHIR) as well as support from the VA's Office of Information Technology (OIT).