JAMA Ophthalmology, March 2017
Valikodath et al. performed an observational study to determine whether symptoms reported by patients on an eye symptom questionnaire (ESQ) were consistent with those documented in their electronic medical record (EMR). The authors found substantial discordance in symptom reporting between these methods and noted that many symptoms reported on the ESQ were absent from the EMR.
Data for the following ocular symptoms were collected from ESQs and EMRs of 162 adult patients: blurry vision, pain/discomfort, glare, itching, light sensitivity, burning/stinging, gritty sensation, and redness. Symptom reporting was analyzed using κ statistics and McNemar tests. Disagreement was defined as a negative report or no mention in the EMR of a symptom that had been noted by the patient as being moderate or severe on the ESQ. Logistic regression was applied to investigate whether the probability of disagreement between records correlated with specific patient factors, physician characteristics, or diagnoses.
ESQ and EMR reports of glare, blurry vision, pain/discomfort, and redness were discordant for 48%, 34%, 27%, and 25% of participants, respectively. Agreement was poor to fair (κ range, −0.02 to 0.42). Discordant reporting usually was characterized by a positive symptom report on an ESQ that was not documented in the EMR (Holm-adjusted McNemar, p < .03 for all symptoms except blurry vision). The likelihood of patient-reported blurry vision being documented in the EMR was greater during visits by new patients than by returning patients. Discordant reporting of blurry vision, pain/discomfort, or redness did not correlate significantly with patient age, gender, diagnosis, or physician characteristics (e.g., duration of practice, volume of work, presence of medical scribe). Whenever a patient reported 3 or more symptoms on the ESQ, the 2 records were not fully concordant.
The authors concluded that because of substantial inconsistencies between patient and EMR reports of ocular symptoms, EMR content should not be considered a comprehensive resource for patient care or large-scale research. Additional studies are warranted to determine reasons for the discordant reporting. (Also see related commentary by Christina Y. Weng, MD, MBA, in the same issue.)
The original article can be found here.