Data collection, EHR and biomedical research

Bates et cols have just published a very interesting time-motion analysis of EHR for specialty clinics in the latest issues of JAMIA (Lo, 2007). The study is interesting in that it uses time-motion analysis, one of the most sophisticated methods for analysis of workflow, to conclude that specialty EHR does not save any time when compared to paper-based systems. From the article, "following EHR implementation, the average adjusted total time spent per patient across all specialties increased slightly but not significantly (Delta=0.94 min., p=0.83) from 28.8 (SE=3.6) to 29.8 (SE=3.6) min." The question we should be asking now is why it does not save time, and the possible answers seem to follow in several categories. First, there might be a mismatch between workflow and the sequence of activities in the EHR. Second, the workflow in existence prior to the EHR might not be appropriately mapped and then reegineered in the EHR. Third, there might be usability issues within the EHR that prevent a gain in time -- these issues could be related to problems in the interface itself or imbalances between free text and discrete variables for data entry. Finally, staff might simply not have reached the appropriate level of training and familiarity with the system to achieve a mature level of speed. These factors are important in that an increase in time is crucial in determining EHR adoption rates but also in terms of additional costs incurred by increased use of time from staff involved in patient care. From a biomedical research perspective, EHR is already playing an essential role in the acquisition of research-related data, and a decrease in the efficiency of our clinics can significantly hamper its progress.