Getting an accurate drug dose into a critically ill or injured child is a complex process that is highly vulnerable to error at each of the steps from weight estimation to drug delivery. As ethical healthcare providers, we need to be passionate about paediatric patient safety and minimise the potential for patient harm at our hands.
Unfortunately, there is a distinct lack of high-quality evidence to guide our practice in the weight estimation-drug dosing process, primarily because there is little evidence on the consequences of drug dosing errors. There is also poor evidence on the actual doses we should be using for many drugs. In light of this, we need to consider what would constitute a good approach to clinical problem for which there is limited evidence to guide us. In as far as emergency weight estimation is concerned, we could either say that it doesn’t matter how accurately we estimate weight because there is no evidence that accurate estimations affect outcomes, or we could say that we need to be as accurate as possible in order to minimise potential drug dosing errors, because that is the most ethical option.
Fortunately, there is some evidence to support my conviction that we need to estimate weight as accurately as possible. There is recent preliminary evidence that at least one-third of clinically stable children who have a weight estimation error of >10% will have a subsequent significant medication error . Scarily, this is likely to be far, far higher and much more serious in children receiving any form of resuscitative treatment: medication errors during emergency treatment may be as much as 39 times more likely to result in harm and 51 times more likely to result in death than non-emergency related medication errors . If this data is to be believed, then this is clearly an important problem that needs to be addressed.
Something else to consider is that when we consider errors in weight-based drug dosing, weight estimation errors are not the only errors that we need to contend with. The effects of cumulative, or compounded, errors must be considered, but these have not been studied in great detail. If we accept that a final 20% error for a drug with a narrow therapeutic index is acceptable, this means that all the component contributors to error must fall within this limit, not just the weight estimation error. The errors which we must consider include the weight estimation error, the errors incurred during the preparation and administration of the medication as well as the potential error in concentrations of the medication itself (governed by statutory regulations). The degree of accuracy that most weight estimation systems achieve has been reasonably well established, while the errors during medication preparation and administration may equal or exceed the weight estimation error . Drug concentration errors may add an additional 10 to 15% error. Thus, a 20% weight estimation error, plus a 20% administration error, plus a 10% concentration error can escalate to a potential 50% error very quickly (yes, adding them gives a good estimation – see Box 1). And that is just not acceptable.
Potential drug dose errors, using adrenaline as an example. The “low” indicates the maximum potential underdosing and the “high” the maximum potential overdosing at each step of the compounded error.
*An adrenaline/epinephrine solution must contain 90% to 115% of the labelled amount to meet United States Pharmacopeia standards (Epinephrine injection. The United States Pharmacopeia: The National Formulary. United States Pharmacopeial Convention, Rockville, MD; 2013).
So, what is the solution? We need to use accurate weight estimation systems and ensure appropriate, goal-directed training in their use as well as in the preparation and administration of emergency medications. Errors must be minimised at every step of the drug dosing-delivery process.
And what about the decision about which weight estimation systems are most appropriate for clinical use? There are three important considerations when evaluating weight estimation systems: their accuracy, their usability and their ability to integrate with a drug dosing system (see Box 2).
The accuracy is relatively easy at first glance: we want the most accurate weight estimation system possible, assuming that it is not prohibitively expensive. However, with the explosion of childhood obesity, we ideally want a system that can estimate both total body weight as well as ideal body weight. This allows us the flexibility to optimise drug dosing for each drug in each patient. Although the outcome data for incorrect dosing in obese children is limited, there is enough to suggest that this is a valid concern. I think that we need to set a target of 95% of weight estimates within 20% of actual weight as a minimum accuracy for any system that we use. At this stage, only the PAWPER tape, the Mercy method and perhaps parental estimates achieve this standard .
The usability of weight estimation systems is also crucially important. Usability relates to how easy a system is to use (ease-of-use), but also how vulnerable it is to human factor errors (resulting from user errors) and patient factor errors (resulting from non-ideal patient position and poor cooperation). The usability of a system will clearly have an impact on its accuracy as well. The balance between usability and accuracy is something important to think about, because it is important to limit processes that substantially add to cognitive load that is already high during paediatric emergencies. Ideally the weight estimation
system should aid in reducing overall cognitive load. The usability factor is something that we have considered carefully in the design of the PAWPER tape. It is designed to be quick and easy to use without adding excessively to the cognitive burden. The Mercy method is a little more complex to use and has a greater vulnerability to both human and patient factor errors .
The ability of a weight estimation system to contribute to the accuracy of downstream processes needs to be considered. This refers to how it can improve the accuracy of drug dosing by how well the system integrates with a drug dosing guide. The best example (in a good way) is an App that can generate an accurate estimate of weight which is automatically used for drug dosing calculations with limited further user input. The worst system is one in which there is no integration at all, such as parental estimates of weight or age-based formulas. Colour-coded systems and other length-based tapes with precalculated drug doses fall in the middle.
Having identified the most appropriate weight estimation system for your setting, the next step is to use it optimally. There are valid arguments about ensuring the system is maximally usable, with a low cognitive load, but there is no system that is completely cognitively neutral. However, complexity during emergencies is the reason that emergency medicine specialists exist. Our training and learning need to prepare us to practise effectively during emergencies and the fact that treatment takes place during an emergency should not excuse a diminished quality of care. These circumstances should not excuse potentially harmful practices.
Last year, I cycled the Cape Town Cycle Tour in South Africa. It is a 108 km route winding through some of the iconic passes in the area. My training for the event consisted of running a 5 km course several times a week for the few months before the event. I finished the cycle race but didn’t do particularly well, which was not too surprising, really. This year, I trained by doing a lot of cycling and, despite adverse weather conditions during the race, I did a lot better than last year. Again, not particularly surprising. The relevance of this story is that appropriate training is important, not just any training!
Let’s assume for a moment that the direct laryngoscopy (DL) vs. video laryngoscopy (VL) debate is over and VL is clearly superior (which it is, right?). VL is perhaps slightly more complex to use than DL. Would we then say that we shouldn’t use it, or would we say that we need to be trained appropriately and practise until we are competent?
It is self-evident that we need to train appropriately for the tools that we use. Often the problem is more becoming aware of this, rather than implementing it. Incorporating weight estimation training and drug administration into routine paediatric simulation exercises will go a long way towards improving paediatric patient safety.
- Weight estimation systems must be evaluated according to their accuracy, usability and ability to be integrated with a drug dosing guide.
- The cumulative medication errors resulting from weight estimation errors and drug preparation and administration can be significant and have a high potential to cause patient harm.
- Appropriate training in the use of weight estimation systems and emergency drug preparation and administration is essential. The incorporation of this practice into simulation training is likewise essential.
- Healthcare providers who specialise in providing care during emergencies must be competent at managing the cognitive loads experienced during emergency care. This can be achieved through appropriate teaching and training. The emergent nature of the presentation should not excuse a diminished quality of care or the use of inaccurate or inappropriate adjuncts (including weight estimation tools).
- Some degree of complexity is inevitable in many aspects of emergency care and adds to the cognitive load. The answer is to limit the complexity as much as possible without negatively affecting patient care and ensure that training is adequate to reduce the effects of the cognitive load.
 Hirata KM, Kang AH, Ramirez GV, Kimata C, Yamamoto LG. Pediatric weight errors and resultant medication dosing errors in the Emergency Department. Pediatric Emergency Care. 2017. 10.1097/PEC.0000000000001277
 National Academy of Medicine. Preventing medication errors: Quality Chasm Series Washington 2006 [Accessed 7 July 2018]. Available from: https://nationalacademies.org/hmd/reports/2006/preventing-medication-errors-quality-chasm-series.aspx
 Murugan S, Parris P, Wells M. Drug preparation and administration errors during simulated paediatric resuscitations. Archives of Disease in Childhood. 2019;104:444-450. https://doi.org/10.1136/archdischild-2018-315840
 Wells M, Goldstein L, Bentley A. The accuracy of emergency weight estimation systems in children – a systematic review and meta-analysis. International Journal of Emergency Medicine. 2017;10(29):1-43. https://doi.org/10.1186/s12245-017-0156-5. https://doi.org/10.1186/s12245-017-0156-5
 Wells M, Goldstein L, Bentley A. The accuracy of paediatric weight estimation during simulated emergencies: the effects of patient position, patient cooperation and human errors. African Journal of Emergency Medicine. 2018;8(2):43-50. https://doi.org/10.1016/j.afjem.2017.12.003