Who uses what: Investigating EHR data documentation in New York City practices

Background: Since 2001, the percent of office-based physicians using an Electronic Health Record (EHR) has risen from 18% to 78%. EHR data holds promise for quality improvement, surveillance and effectiveness research. Quantifying EHR data quality is an important first step in understanding the potential of EHRs and identifying data limitations. Using aggregate data, the NYC Health Department's Primary Care Information Project (PCIP) explored systematic patterns in EHR documentation. Objective: Identify EHR documentation patterns from 2009 to 2012 among PCIP practices. Methods: For each year from 2009 to 2012, we compared total patients seen to the subset of patients with: a valid BMI in vitals; standard race/ethnicity in demographics; one of the 2,000 most common LOINC codes in lab results; or any prescription in medication management. Practice documentation rates were reported for the first and final years and significance of change ascertained using Wilcoxon signed-rank tests. A total of 278 PCIP practices met the inclusion criteria (≥25 patients seen in 2009). Results: On average, practices documented BMI for 44.1% of patients in 2009. By 2012, that number rose to 72.1%. In 2009, practices recorded race for 30.1% of patients, and ethnicity for 42.8%. Four years later, 47.5% had race and 69.2% had ethnicity. By 2012, an average of 45.0% of patients had ≥1 electronic lab results, as compared to 16.1% in 2009. By 2012, practices had prescribed ≥1 medication to 71.8% of patients on average, as compared to 52.1% in 2009. All changes were significant (p<0.001). Discussion: Data completion varies by area of the EHR and by practice. EHR use is improving over time, incentivized by Meaningful Use. Reimbursements should focus on improving documentation. Building indicators using practices that document well in a given section of the record may improve data for research and surveillance.