I have previously posted concerns regarding the imminent introduction of electronic prescribing to hospitals and their potential for causing more problems than they are worth for ED’s in particular if they do not have the buy-in of clinicians, adequate computer access and efficient user interfaces – see here and here.
This month’s EMA journal (Emergency Medicine Australasia (2009) 21,373-378) has a paper published by Fernando et al discussing the implementation of electronic order entry for pathology and radiology in a Sydney hospital in late 2006.
The paper suggests a rather scathing criticism of the inadequate change management, the shift in data entry chores from laboratory staff to ED doctors when they can least afford to have more computer work to take them away from direct patient care when they have to contend with the stress of bed access block, but perhaps the greatest frustration appears to be directed at the inefficient and sometimes dysfunctional user interface of the software itself and the “can’t do” attitude of the software vendor in response to requests for change.
In other words, the new system took MORE time for ED doctors to order investigations than the previous paper-based system, as they had to find a computer, log in, search for the patient, search through a number of pre-formatted boxes and layered screens to find the test they wanted to order. “One doctor noted it took more than 50 key strokes to order basic blood tests and an x-ray for a patient with hip fracture, whereas a paper-based system would have required only 2 pieces of paper and some brief hand written notes”.
The system also had dysfunctional mandatory requirements such as requiring you enter a pregnancy status even if the patient was male, but it was the repetitive and time consuming features that were the main source of aggravation to ED doctors.
They noted a “high level of frustration” about the inadequacy of ED involvement in the implementation process. Perhaps the lack of involvement is not the fault of Cerner as it would not be surprising given that most ED staff have very little spare time if any to participate in such processes.
They also noted that widespread clinical opposition to this same software resulted in its removal from the Cedars Sinai Health System in LA in 2003.
This report should be sending waves of concerns to both hospital and government administrators who are looking to roll out such systems.
Perhaps the vendors need to talk more with clinicians and find out how we work and what we need to be more efficient, not less efficient. The same issue has occurred with Victoria’s PMI system which is less efficient for clerical staff than the previous DOS-based system.
Why are the software vendors ignoring the needs of users?
To help them out, I have just posted screen shots of my version of an Emergency Department Information System (EDIS) which is designed from a clinician’s point of view and places patient safety and clinical efficiency as it’s prime priorities – see here for the screen shots, and note guys all the speed buttons and in-built intelligence, so for repetitive tasks, you only hit one button to save 10-20 key strokes – it’s not that hard, but it does involve understanding what clinicians do – and that takes a bit of work!
If vendors are really serious about making such systems work then they have to give the users some wins.
A feature that is desperately needed in hospitals far more urgently than electronic order entry, but related to it, is a hospital-wide, efficient and effective investigation results checking system which works with clinicians not against them.
Without such a system, doctors become much more inefficent as they need to manually follow up investigation reports and then with inadequate information decide how to deal with the abnormal results which leaves patients at risk of un-managed abnormal results.
Furthermore, electronic order entry runs the risk of doctors making mistakes more rapidly by making it easier to order investigations for the wrong patient on the computer because there is no visual feedback of a patient photo to prompt them and busy doctors not infrequently select the wrong patient from a computer list and don’t read everything before they start typing.
We need to ensure these systems ARE more efficient than what we currently do otherwise they are not worth having.
