A study published in the Journal of the American Medical Informatics Association (JAMIA) may be one swallow and not the spring, but points to something doctors have been reporting anecdotally for years. Researchers examined initial progress notes of patients admitted to Beaumont Hospital in Royal Oak, Michigan both before and after the Epic Systems EHR implementation (POLITICO Morning eHealth) in 2012. Their sample of 500 notes examined five specific diagnoses with invariable physical findings: permanent atrial fibrillation, aortic stenosis, intubation, lower limb amputation and cerebrovascular accident with hemiparesis. The error rate of EHRs compared to the paper charts was 24.4 percent versus 4.4 percent. Residents were better at EHR-ing than the more experienced attending physicians for inaccuracies (5.3 percent v. 17.3 percent) and omissions (16.8 percent v. 33.9 percent). As this is an older snapshot, it may have narrowed with familiarity and training, but this is in line with prior reporting in multiple countries (here) that customization by real clinicians needs to be part of the implementation (designed by IT people without clinical background), often design doesn’t meet clinical needs, many have glitches and that they take entirely too long to fill out, notoriously in mental health (see JAMIA study from April). And let’s not get into the plagues of hacking, ransomware and health data exchange. HealthcareITNews, JAMIA (abstract only)