Introduction: A previous Canadian emergency department (ED) model determined predictors of increased workload using a manual chart review to elucidate comorbidities. We designed an electronic algorithm to capture all comorbidities based on the Charlson Comorbidity Index (CCI) for a 5 year period preceding the ED visit from the regional inpatient database. Our objective was to identify predictors correlating with physician time require to treat patients and thus develop a multivariable model to predict physician workload. Methods: From May to September 2015, two research assistants (RAs) shadowed a random sample of physicians from the six urban EDs in a single health region. They documented time spent performing clinical and non-clinical functions for patient visits. A linkage with the previously validated regional ED database was used to obtain triage acuity, age, gender, mode of arrival, and CCI scores. Multiple linear regression was used to describe the associations between predictor variables and total physician time per patient visit as well as time spent on history and physical exam and to derive an equation for physician workload. RA inter-rater reliability was assessed on 107 MD-patient interactions. Results: Over the 4-month period, 873 patient encounters were documented. Data from 599 completed encounters were included in the model. The median age was 49.4 (SD 22.8) and 52.2% were female. Overall, 16.0% were admitted to hospital, 64.9% of patients were CTAS 1-3, 19.6% of patients arrived by ambulance, and 15.5% of patients had a CCI score of ≥ 1. The mean time spent on history and physical was 7.0 minutes (SD 4.73) and mean total time was 19.4 minutes (SD 11.6). Using a linear regression model with total time as the dependent and EMS arrival, CTAS, and age as the independent variables, having any CCI score is a significant predictor of total time (p = 0.03). with a difference of 2.9 minutes between CCI positive versus negative patients. Higher acuity was the most significant factor associated with time spent with a mean difference of 4.4 minutes per CTAS category. The intraclass correlation coefficient value was 0.99 (95% CI 0.97-1.00) indicating excellent reliability. Conclusion: The electronically derived CCI does have value in the development of a physician workload model and can replace the use of manual chart review to define patient comorbidities.