Leveraging electronic health records to support chronic disease management: The need for temporal data views

Electronic health records (EHRs) may improve primary care management and care co-ordination. However, studies have failed to consistently show improved quality with EHRs in chronic disease care. For example, diabetes disease management programmes using physician reminders, performance feedback and nursedriven structured care have been shown to improve process outcomes such as rates of haemoglobin A1c testing, but not intermediate clinical outcomes, such as glycaemic control [5] Recent reviews show positive results fromabout two-thirds of trials of clinical decision support, but absolute changes in clinical outcomes are small; a Cochrane systematic review of point-of-care reminders for hypertension management reported a median reduction in systolic blood pressure of only 1.0 mmHg in intervention groups. [6-9] Finally, some studies have shown that EHRs have failed to improve care when implemented as part of a disease management programme, when clinical decision support is incorporated, or when used for care co-ordination across settings. [10-13] EHRs may have failed to improve chronic disease care in these studies in part because most EHRs in use today do not facilitate longitudinal data management. [14] We believe that the contribution of health information technology has been limited because EHRs are largely visit oriented and clinicians need contextual information for each data point. For example, a provider treating a patient for hypertension needs not only the medication list, but also context about how the patient’s blood pressure has responded to each medication that the patient has tried in the past. [15] To attempt to provide context, systems designed for clinical care have attempted to display all patient data on one screen, but this causes cognitive overload from the presence of extraneous data. [16] Chronic disease management requires filtering pertinent diagnosisoriented information. For this reason, we propose the term ‘temporal data views’ to mean both an improved interface and a filtered view of patient data. Chronic disease management requires numerical data, best displayed using visualisation, as well as nonnumerical data, best displayed using diagnosis-oriented text summaries. Neither data visualisation nor diagnosisoriented text summaries have been widely adopted into the EHRs commonly used in primary care practices today.