Scalability of real-time data delivery from electronic health records

Electronic health records (EHR) are an essential component of IT in healthcare. Low delivery latency (DL) of real-time data (RTD) from EHRs is crucial for other systems and applications. Scalability has a direct impact on latency and the two are closely related. The methods for RTD delivery include REST hook, WebSocket, reverse proxies and database triggers. The objective of this work is to test the ability of the RTD delivery methods to maintain low DLs when increasing the number of recipients.

Four test setups were created, consisting of two machines connected in a local network via cable, to record the DLs for the 4 methods. The tests were performed with 5, 10, 15, 20, 30, 40, 50, 60, 70 and 80 recipients.

The results indicated that the most scalable method was DB trigger whose mean DL rose only from 18.02 ms to 27.81 ms when the number of clients increased from 5 to 80. Second best mean DL was that of the reverse proxy which rose from 20.38 ms to 34.61 ms. The mean DL of REST hook increased steadily until 40 recipients and thereafter the values spiked to 17817.16 ms. WebSocket showed consistent performance as the number of clients increased.

Our findings indicate that database triggers are the most effective method for implementing RTD delivery. This highlights the urgent need to prioritize the integration of real-time databases with EHR APIs, which can ultimately lead to more responsive healthcare systems, and thus to improved patient outcomes and a higher standard of care.