Unscripted Errors – What a Beat-Up!

Over the last few years the Australian Newspaper has developed a reputation for quality reporting in the e-Health space, led largely by Karen Dearne and James Riley (who has been doing good work on the “Access Card” front). It is therefore quite disappointing to note the article, entitled Unscripted Errors, which appeared today providing such alarmist and ill-structured comments from a different Australian journalist.

In essence the article argues that electronic prescribing by GPs, in the absence of them using the software as it was designed and the software lacking the key features (electronic decision support) that make it safe will not help reduce prescribing errors. To suggest this is stating the ‘bleeding obvious’ is a monumental understatement – and indeed is confirmed by the hospital study cited in the article.

It is little more than common sense to say that safety requires all parts of a system to work properly. (Thus having perfectly sound wings on an aeroplane does not help avoid a crash when the engines fail!). It is no news that poor computer systems in the hands of the untrained or careless may do more harm than good. It is also true to say that the current generation of Australian GP prescribing systems, used carefully and intelligently by clinicians trained in their use, are very safe and can make a real difference

In the case of prescribing the ideal system to do the job will have the following attributes:

1. It will be part of a comprehensive Electronic Patient Record which captures relevant information (ideally in properly coded form) regarding the patients illnesses, allergies and other medications (including non-prescribed ones).

2. It will have access to a rich electronic decision support framework which provides guidance as to prescribing based on age, height, weight, disease, other medications and allergies at the least.

3. All the alerts and recommendations made by the system will be evidence based and properly researched.

4. The decision support will interact with the clinician at the point of decision making to ensure all relevant information is considered.

5. The system will be easy to use and have its data bases of knowledge regularly updated and refined.

However even all this is not enough, only when the clinician is both properly trained and prepared both t0 enter all relevant information and respond intelligently to the decision support warnings and alerts will the substantial beneficial reduction in errors and patient harm be achieved.

It is what is needed, rather than what can go wrong in the hands of clinicians given poor software with out proper training, that should have been the emphasis of the article. As it stands the article did nothing but alarm, rather than fostering an effort at improvement which is genuinely needed.

The evidence that such systems work is now overwhelming if they are properly implemented.

Getting this sort of software onto clinician’s desks is yet another area where national leadership could save both money and lives. We should be working to have proper systems, properly implemented, put in the hands of our clinicians, not worrying ourselves into paralysed inactivity because a few might misuse older ones.

David.