AI and Machine learning

The need for better linkage between administrative data and clinical datasets

The increased use of the National Health Institute Research Database (NHIRD) by the Taiwanese research community has highlighted the fundamental importance of an accurate and timely national health database in monitoring the state of a health system.1 However, the NHIRD shares the common limitation of other administrative datasets—it lacks essential clinical outcomes such as glycated haemoglobin and blood pressure measurements, which are particularly useful for looking at the effects of policy on intermediate outcomes in patients with chronic conditions such as diabetes.2

One way to overcome this limitation is to establish better linkage between administrative data and clinical datasets. This has been done in England, for example, by linking the administrative hospital episode statistics database with the much richer data in electronic primary care records. The English experience also shows that such linkages need support from the public and patients—otherwise they run the risk of experiencing the same problems as the care.data programme.