Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Lesson of the week re-interpreted
- Roger L Weeks (22 April 2007)
-
Could I please support Dr Roger Weeks 100%?
- Jim Page (23 April 2007)
-
Simple Sums, Not The Meaning of the Universe
- William T Stevenson (24 April 2007)
-
RE: information technology cannot guarantee patient safety.
- Saskia N. de Wildt, Ron Verzijden, John N. van den Anker, Matthijs de Hoog (8 May 2007)
|
|
|||
|
Roger L Weeks, GP Deanhill Surgery SW14 7DF
Send response to journal:
|
de Wild et al are right to highlight the dangers of dodgy software putting patient safety in peril. However the conclusion inherent in the title of this weeks lesson –“Information technology cannot guarantee patient safety” goes too far. Properly written knowledge databases and software could go a long way to facilitate much greater levels of patient safety particularly in the dangerous and fraught situation described. The criticism in the article concentrates too much on the shortcomings of Excel but ignores other more worrying inadequacies. I believe that use of Excel or other spreadsheets for dose checking (with or without the locking facility) is totally inappropriate particularly if they make users key in lots of data and allow it to be put into the wrong field. The problem, however, is not just data going in the wrong place - it is essential that the concepts of strength and dose are not confused which seems to have occurred in this case and that every entry is clear, follows accepted standards and its purpose is clear. This is not the case here: Worrying and wrong detail seen in the tables: There are non standardised drug names e.g. • NaBic (presumably sodium bicarbonate not a Biro) • Norepinephrine (standard name Noradrenaline), There are non standardized strengths e.g. • Atropine 1ml=5mg (rather than the standard 5mg/mL) • Epinephrine (1:10000) and our old friend • NaBic 4.2% and • Glucose 20% There are no strengths in Dopamine, Norepinephrine (standard name Noradrenaline) Dobutamine, Ca Sandoz (how did this non-generic creep in?), Midazolam, Phenobarbital and Phenytoin. In the section under dose/kg in infused drugs there is more confusion – for Norepinephrine we find “0.1ug/kg/min” – the units here are micrograms – the character ‘u’ should not be used instead of ‘μ’ and if this character is not available one should resort to the standard mcg. In addition it is not clear from the snapshot what the recommended flow rate would be (and should not this be in drops per minute?) It is perfectly straightforward to write robust software for handling dose calculations but this is not a job just for the computer programmer. The first requirement is to assemble the ‘knowledge domain’ of the application– in this case all drug products (and all their details in standardised format) that would ever be needed in paediatrics and the medical information domain for their use (indications, contraindications, side effects, interactions dosage and so on) Next the knowledge concepts and related terminology need to be organised, preferably in to some form of hierarchical thesaurus and databased using a database programmer. Now the algorithm to do the dose checking can be written and checked. Finally the programmer can write his program. This program makes calls to the knowledge domain database on he basis of the strictly controlled entries of the user. The drugs required are selected from the database ensuring real products are chosen by the user as are the weight, age and medical conditions. The knowledge database must contain all the drugs and other treatments including strengths, formulations, dose per kilogram body weight, dose for specific indications, routes and rates of administrations. The whole job could be done by a team of one doctor or pharmacist with the necessary knowledge (or at least access to it) one database expert and one programmer – in one year or in standardised units three man years. Competing interests: Roger Weeks is the managing director of SafeScript Ltd. and co-author of its product: the World Standard Drug Database |
|||
|
|
|||
|
Jim Page, Dental Practitioner and Dental Tutor TN4 8BH
Send response to journal:
|
Could I please support Dr Roger Weeks 100%? I am a mere amateur user of computers but I have been trying to use them for over 20 years and have found out how easy it is to make errors if one uses the wrong type of program So many times I have observed people using spreadsheets in totally inapropriate situations. I know that Excel is a very clever application and it may be possible for people cleverer than me to use it without problems but I have made so many mistakes myself and seen so many others make mistakes using it for fairly simple calculations that I do not think that it should be used in "mission critical" situations where there is not time to double or triple check the calculatons. A relational database is inherently much safer type of program for this type of application and is what should be used for "mission critical" calculations as made clear by Dr Weeks Competing interests: None declared |
|||
|
|
|||
|
William T Stevenson, Consultant Radiologist Royal Lancaster Infirmary LA1 4RP
Send response to journal:
|
Pretty soon we can expect a Lesson of the Week emphasizing that it's a bad idea to remove the wrong kidney. The point here is that if a bunch of zombies (presumably members of the group who believe that information from the Internet must be reliable, that walking along the street engrossed in a mobile phone is an activity worthy of a sentient being and that counselling and mentoring are good ideas) are permitted to entrust all responsibility for a simple calculation with a potential deadly effect to a calculator, things are apt to go wrong. It is a truth universally acknowledged that people who operate on hearts know which way the blood goes round; people who wish to inject dangerous drugs are expected to be able to find out how much to inject. They are not permitted to claim "I was never any good with figures"-if that's too tough they need to look for a job in Ethics or something. You take a piece of paper, round the numbers to integers, get the powers of 10 correct, and get the rough answer. Then see if it's reasonable. Then use the calculator, spreadsheet or supercomputer, if you can't rid yourself of the delusion that the more decimal places the better. Surely the combination of A-levels, years at medical school and MMC can achieve this without pleading that we are helpless victims of those terrible computers. Isn't the Public entitled to expect that a little thought has gone into 'fail-safing' these systems, in view of the inevitable failure of components? I suggest a little notebook with writing in it. Competing interests: None declared |
|||
|
|
|||
|
Saskia N. de Wildt, Clinical Fellow Pediatric Intensive Care and Clinical Pharmacology Hospital for Sick Children, Toronto, ON, Canada, M5G 1X8, Ron Verzijden, John N. van den Anker, Matthijs de Hoog
Send response to journal:
|
We thank drs Weeks, Stevenson and Page for their responses to our article “Information technology cannot guarantee patient safety”. We fully agree with dr Weeks, that Excel is not the most appropriate program for dosing calculation and dose checking. Indeed, other more appropriate software applications are available. As dr Page correctly points out, a notepad should be sufficient to make calculations. One of the arguments used by our PICU staff to not use a computerized resuscitation sheet so far, but instead, ask residents to do all 12 calculations by hand, was to familiarize these young doctors with drug doses for different age groups. Ultimately, this is shown to result in less mistakes in drug calculations. However, these young doctors work long and extremely busy hours, in itself shown to be associated with an increased risk of mistakes. Hence, used as they are to new technologies as PDAs (personal digital assistants), Blackberries, internet, etc, it is imperative that they try to use these technologies to help them do their job. To date, despite huge efforts for evidence-based guidelines, drug names, doses and formulations are not uniform across hospitals, let alone internationally. Consequently, there remains a need for individualised software applications based on local protocols. Given the considerable amount of time and resources to develop these applications it can take months to years before they are available. In the meantime, it is obvious that individual doctors will continue to use ‘hand-made’ non-validated applications. Ignoring, disapproving or ridiculizing this use, will not improve patient safety, it is our Lesson of the Week, to make doctors aware of the inherent pitfalls. Competing interests: None declared |
|||