Jason (his name has been changed to protect his identity), an 18-month-old baby, was taken by his parents to a hospital in the U.S. with a fever. At the time the doctor had him checked in the emergency room, the baby’s temperature was 42º C (107.6º F).
After reviewing some studies, the doctor asked Jason’s parents to keep him in the hospital for observation and used the Electronic Health Record (EHR) to prescribe acetaminophen, commonly known as Tylenol or paracetamol, in order to lower the fever.
24 hours later Jason was in a coma, with his liver severely damaged.
The doctor had prescribed a dose of acetaminophen twice the appropriate for a baby like Jason. In that hospital’s EHR, the pediatric prescription requires the doctor to select the drug and its power or concentration, enter the doses in milligrams per kilogram and the patient’s weight in kilograms. Jason’s weight was 12.25 kg (27 lbs). However, the doctor entered 27 kilograms (60 lbs) in the EHR. He didn’t realize he had entered the child’s weight in kg, so he entered a value that more than doubled the correct one. Later, the nurse gave the baby a double dose of acetaminophen for several days, until Jason showed symptoms of liver intoxication.
Thus, the question here is: who is responsible for Jason’s error-induced disease? We could easily blame the doctor for his mistake, since he should have known that healthcare systems use the metric system. Moreover, he should have seen the input in the screen prompted to enter the weight in kilograms. We could also blame the nurses: they should have realized that Jason was not a 60 lbs kid and that the prescribed dose was way too much paracetamol for a baby.
To answer this question, it is important to highlight that using an EHR is not the goal of doctors and nurses. An EHR is useful for registering patient’s data and communicating orders, but their main purpose is to take care of patients (i.e. detecting symptoms, making decisions, and providing treatments). Furthermore, an emergency room is a chaotic environment, where it is difficult to slow down to take into account nuances like the unit of weight for an input box.
This medical error leads us to reflect on how better design could prevent similar situations in the future. In the following sections, we are going to analyse this case from a problem-solving point of view, looking for solutions based on user-centered design (UCD) principles.
A well established UCD principle states that apps should speak the users’ language. In the previously presented case, measuring things using the users’ units. But what can we do in the case of the US? We are all probably aware that the US is one of the few countries in the world that has not officially adopted the International System of Units (SI or metric system). This is despite the fact that experts recommend the metric system within medical environments for standardization. Accordingly, the EHR used in the hospital that treated Jason was configured to enter the patient’s weight in kilograms. But, is this safe? Which is the US doctors and nurses’ system of measurement? Is it the metric one, or their everyday customary system, which is shared with their patients? When a doctor or a nurse asks about a baby’s weight, their parents are bound to express it in pounds. There are hospitals that still use scales that weigh in pounds, or are configured to show the weight in pounds.
Another key rule of user-centered design says that equal elements should look and work the same way. If there are elements that look alike, users expect them to mean and behave the same. This is valid for everything, from objects to apps, from urban to interactive design. Consistency applies to colors, typography, shapes, links, icons, menus, words, behaviors, and units of measure.
Any variation from standards and customs has the risk of being misunderstood. In the US for example, an EHR input of weight in kilograms is consistent with other healthcare software, forms and terminology. However, it may be inconsistent with common family records.
Any innovative design will have to be learned by users, based on their previous knowledge. On the one hand, consistent design is intuitive as it relies heavily on previous experiences and encourages learning. Inconsistent design, on the other hand, is hard to use and learn because it is counterintuitive, users are forced to constantly think about how a particular system works.
For users, if an app doesn’t behave like expected, it could lead to an inadvertent error. This is especially important in healthcare software, since errors could lead to life-threatening events. Given this situation, it is always convenient to design with human error in mind.
Although, we usually think that efficiency of use (reduction of the time and effort needed to perform tasks) is the most important usability value, that is not always the case. In healthcare, time is critical, but patient safety is even more important. Every designer, especially those working with critical systems (e.g. aircraft navigation, stock market, nuclear plant, or sensitive data), have to take into account the application of Murphy’s Law to users’ behavior: if users can make a mistake, they will make it. They will commit it in the worst possible situation. And they will not be aware of it!
Therefore, one step towards safety is to provide feedback, so users take notice of what they are doing. Take Figure 2 as an example, after the input in kilograms, the app can show the equivalent value in pounds.
Nevertheless, this design does not ensure the users will notice that the weight is expected in kilograms, or that they will read the provided value in pounds. So, we cannot assure they will be aware of having committed a mistake.
Doctors and nurses may talk between themselves in terms of pounds, and then convert the weight to kilograms just when they are registering it. Hence, one possible alternative is to make the unit of measurement selectable. Users could choose whichever unit they manage, or leave the default one. Kilograms may be the default, since that is the official standard, see Figure 3.
Still, users may ignore the default option for unit of measure. For this reason, we could ask the users to type the unit symbol. By making this option explicit, users would be more conscious of the values they enter. Now we should have to validate the typed unit and check it against a list of legal options. Even as this alternative has the disadvantage of making the process a bit slower, that is a cost we can afford in order to ensure the users are fully mindful of the values they enter.
Alarms are a great tool to avoid risky situations for patients. However, too many alarms can cause fatigue among doctors and nurses, until they are finally ignored. They should be designed as a balanced system that prioritizes life-threatening actions over mild risks.
The EHR could detect the entering of an outlier figure. Weight can be checked in concordance with the patient’s age and length, based on statistics for the population. According to CDC, a male, 18 month-old baby, could weigh between 8.75 to 13.75 kg (19-30 lbs).
CDC’s chart of weight-per-age percentiles for 0-24 months boys showing
the expected range of weight for an 18-month old baby.
So, even if Jason was a really big baby, at 27 kg he would have been on the range of a 6 to 11 year-old child!
CDC’s chart of weight-per-age percentiles for 2-20 years boys, showing
the expected range of age for a baby weighing 27 kg (60 lbs).
If the system could check the probability of a value, and finds it very low, it could prompt an alarm. And thus, avoid a possible life threatening situation.
Finally, the total dose could be checked if it is correct for the age. An 18-month-old baby should not take more than 5 ml of 160 ml/ 5ml acetaminophen syrup. Jason took 7.5 ml every 6 hours. This check could prompt an alarm when prescribing and administering medication.
In this article we have reviewed an example of poor usability of EHRs that can lead to errors in administering medication, and jeopardizing patient health. This kind of case was cited as a frequent occurrence in the use of healthcare software. We have shown how the principles of user-centered design can help to improve the patient’s safety by preventing errors.
This particular case presents a tricky problem, since it is not clear which is the correct system of measurement of choice. We discussed the risks that emerge when there is a difference between the use of the US customary units by ordinary people and metric ones by healthcare professionals. Based on that, we proposed several solutions with the intention to make the user more conscious of their actions: giving feedback in US customary units, providing the possibility to choose the unit, and, finally, forcing the user to enter the unit.
Beyond the interaction design, we proposed a smart system of alerts that actively catch the attention of users in case they enter a risky or anomalous value.
This is just one example of how difficult it can be to design safe healthcare apps, that not only make the jobs of doctors and nurses easier, but enable patients to enjoy better health services.
For more information about the challenges that healthcare IT has to address and why UX should be included in every project concerning healthcare software, check out our whitepaper “UX in healthcare technology: Why is it more relevant than ever?”.
Patients are not the only victims of errors regarding the use of metric or US systems. On September 23, 1999, the Mars Climate Orbiter got burned in the Martian atmosphere. There was a navigational error due to commands from Earth being sent in US units (in this case, pound-seconds) without being converted into the metric standard (Newton-seconds).
December 17 / 2019
Digital technology enables organizations to boost productivity and better engage with customers. However, many technology initiatives continue to see cost overrun, damaging the bottom line and undermining the value...Read post