Automated reporting capabilities remain elusive

The ability to automate reports for quality measures would save providers time and provide a clear view of their progress over time, but that ability is still out of reach, according to researched published in the January issue of Annals of Internal Medicine.

Quality measures typically have been derived through administrative claims data or the manual review of paper charts, but these methods have limitations. “Although administrative claims can generate data for large samples of patients, they lack clinical detail,” wrote lead author Lisa M. Kern, MD, an associate professor of public health at Weill Cornell Medical College. “Manual review can generate clinical detail, but only on small samples of patients because of its time-consuming nature. Automated electronic reports of quality from EHRs can potentially address these limitations, offering clinically detailed data for many patients.”

Despite the potential benefits associated with electronic reporting, much existing literature finds it is not currently feasible, a conclusion also reached by Kern and her colleagues. To test the accuracy of electronic reporting, they created reports for 12 measures on 1,154 patients treated at a federally qualified health center from both manual review and an automated process.

Researchers found “wide measure-by-measure variation in accuracy and statistically significant differences for three measures.”  Electronic reports underestimated rates for measures of asthma medication and pneumococcal vaccination, and overestimated rates for a measure of cholesterol control in diabetic patients.

The problem with electronic reporting is that EHRs aren’t capable of scanning unstructured data fields, as manual reviewers can, and fail to recognize care that was delivered, according to Kern. “For automated reporting to be valid, all of the following must occur: clinicians have to document the care they deliver and maintain the accuracy of data in the EHRs, documentation must be amenable to automated reporting and electronic specifications have to capture the same fields that a references standard manual reviewer would consider.”

While technologies like natural language processing could better allow EHRs to accomplish this in the future, the inaccuracy produced by electronic reports is also the fault of providers failing to use EHRs properly. “It suggests that physicians need to recognize EHRs not as electronic versions of paper records, but as tools that enable the transformation in the way care is delivered, documented, measured and improved,” Kern concluded.