By placing referrals into a queue, we can track each referral individually and assign multiple bots. As referrals are a high-priority and need to be processed in real-time, this approach is essential.
The first stage ensures that MS Word is not open, and there are no residual files left in any of the data folders. While this is not necessary as files are removed at the end of each process run, this is more of a safety feature.
Checks are made to see if the Electronic Referral System (ERS) is running; if not, it is started and logged on.
The next stage searches for the referral to be processed using the Unique Booking Reference Number (UBRN). This data item is used as a key and is ideal for searching audit logs for historical referral queries.
When developing the bot code, it is sometimes necessary to do some crazy things to make up for the inefficiencies of the target application. In this situation, we have to teach the bot to wiggle the mouse to keep the screen alive.
With the referral identified the clinical information screen is opened. Depending on the time of day, system responsiveness varies, and so a check is made to see if the referral information is displayed on the screen before proceeding.
With the clinical information screen open, key clinical data is extracted from the page (NHS Number, Specialty and Referral Date) before downloading the many clinical attachments for merging and sorting.
The next blog will focus on the attachment process - it is complicated and a long one!
All objects and processes for this automation are available for free from the NHS Marketplace.