Page - 42 - in Applied Interdisciplinary Theory in Health Informatics - Knowledge Base for Practitioners

2.1. Case Study 1: The value of using national summary electronic health records For any formal evaluation of an electronic record system (EHR), whether at a single institution or at nation scale, measurements need to be taken at multiple points along the value chain (Figure 4). The outcome at any stage can only be understood by modeling earlier upstream events. Thus, failure to demonstrate clinical outcome changes following the implementation of an EHR might arise because of problems with events early in the chain e.g. record quality. Alternately, a lack of impact on outcomes may be unrelated to the EHR (for example organizational challenges may prevent important information from the EHR being translated into process changes).[12] Figure 4: The information value chain provides a simple causal model connecting EHR use and clinical outcomes. Each step is characterized by different measures, and is dependent on different elements of shared record system design and use (adapted from Bowden and Coiera, 2017). For example, imagine that a government has built a national summary health record for every citizen. The system is classed as a success because a large number of citizens have records created for them, and there is a regular stream of record updates every month. What if we however look not at how much data are uploaded into the system, but how often clinicians queried the data? If the system was not often used to support clinical care, perhaps the evaluation might be very different. Evaluation might reveal that the system was not easy to use by clinicians (who therefore were abandoning it), or that the information within the records was not useful, or even that the systems in place to access the records were not mature compared to the data upload arm of the system. Finally one might look at the downstream impact of system use on the cost and quality of care delivered. What changes to care result from accessing the record? Do these changes translate into better decisions that improve patient outcomes or create service efficiencies? It might prove very difficult for a government to answer these final questions, and very easy to provide data about record or usage numbers. There is however no logical reason to assume that usage of a system translates into changes in end outcomes. E.Coiera /AssessingTechnologySuccessandFailureUsing InformationValueChainTheory42